Kevin W. Bowyer - Publications.

• Introduction to the Special Issue on Recent Advances In Biometrics,
  Kevin W. Bowyer,
  IEEE Transactions on Systems, Man and Cybernetics - Part A, 40 (1), May 2010, to appear.
  pdf of this paper.
  As with BTAS 07, a biometrics-themed special issue of Systems Man and Cybernetics - Part A was organized following BTAS 08. However, unlike the SMC-A special section drawn from BTAS 07, this special issue had an "open" call for papers, meaning that submissions were not limited to papers presented at BTAS 08. ... A total of 31 submissions were received for the special issue. ... The ten papers that appear in this special issue represent the result of this process. Four of the ten papers appearing in this special issue are revised and extended versions of papers presented in the closing session of the BTAS 08 conference, a session which, by BTAS tradition, is reserved for the submissions that receive the overall best reviews from the conference program committee. The papers in this special issue cover a range of different biometric modalities, including face, ear, iris, signature and multi-modal. In the area of face recognition, there are papers dealing with 2D, 3D, and hand-drawn sketches. Also, the papers in this special issue range from relatively application oriented to relatively theoretical. One common theme is that each paper addresses an important current topic in biometrics research and makes a novel contribution to the state of the art.
  

• Factors That Degrade the Match Distribution In Iris Biometrics,
  Kevin W. Bowyer, Sarah E. Baker, Amanda Hentz, Karen Hollingsworth, Tanya Peters and Patrick J. Flynn,
  Identity in the Information Society.
  DOI link. (open access)
  We consider three "accepted truths" about iris biometrics, involving pupil dilation, contact lenses and template aging. We also consider a relatively ignored issue that may arise in system interoperability. Experimental results from our laboratory demonstrate that the three accepted truths are not entirely true, and also that interoperability can involve subtle performance degradation. All four of these problems affect primarily the stability of the match, or authentic, distribution of template comparison scores rather than the non-match, or imposter, distribution of scores. In this sense, these results confirm the security of iris biometrics in an identity verification scenario. We consider how these problems affect the usability and security of iris biometrics in large-scale applications, and suggest possible remedies.
  

• Iris Recognition Using Signal-level Fusion of Frames from Video,
  Karen P. Hollingsworth, Tanya Peters, Kevin W. Bowyer and Patrick J. Flynn,
  IEEE Transactions on Information Forensics and Security 4 (4), 837-848, December 2009.
  pdf of this paper.
  DOI link.
  We take advantage of the temporal continuity in an iris video to improve matching performance using signal-level fusion. From multiple frames of a frontal iris video, we create a single average image. ... No published prior work has shown any advantage of the use of video over still images in iris biometrics.
  

• Using Fragile Bit Coincidence to Improve Iris Recognition,
  Karen P. Hollingsworth, Kevin W. Bowyer and Patrick J. Flynn,
  Biometrics: Theory, Applications and Systems (BTAS 09), September 2009, Washington, DC.
  pdf of this paper.
  ... Previous research has shown that iris recognition performance can be improved by making these fragile bits. ... We find that the locations of fragile bits tend to be consistent across different iris codes of the same eye. We present a metrics, called the fragile bit distance, which quantitatively measures the coincidence of the fragile bit patterns in two iris codes. We find that score-fusion of fragile bit distance and Hamming distance works better for recognition than Hamming distance alone. This is the first and only work that we are aware of to use the coincidence of fragile bit locations to improve the accuracy of matches.
  

• Contact Lenses: Handle With Care for Iris Recognition,
  Sarah Baker, Amanda Hentz, Kevin W. Bowyer and Patrick J. Flynn,
  Biometrics: Theory, Applications and Systems (BTAS 09), September 2009, Washington, DC.
  pdf of this paper (not final version).
  Many iris recognition systems operate under the assumption that non-cosmetic contact lenses will not affect match quality and the convenience of the system. In this paper we show results opposing this belief ... The false reject rate varies with different types of contacts and the artifacts they produce on iris images.
  

• Stability of the Iris Match Distribution,
  Presentation only of work done with Karen Hollingsworth, Sarah Baker, Amanda Hentz, Tanya Peters and Patrick J. Flynn,
  Biometrics Consortium Conference (BCC), September 2009, Tampa, FL.
  pdf of slides.
  

• The ND-IRIS-0405 Iris Image Dataset,
  Kevin W. Bowyer and Patrick J. Flynn,
  Notre Dame CVRL Technical Report.
  pdf of this report.
  The Computer Vision Research Lab at the University of Notre Dame began collecting iris images in the spring semester of 2004. The initial data collections used an LG 2200 iris imaging system for image acquisition. Image datasets acquired in 2004-2005 at Notre Dame with this LG 2200 have been used in the ICE 2005 and ICE 2006 iris biometric evaluations. The ICE 2005 iris image dataset has been distributed to over 100 research groups around the world. The purpose of this document is to describe the content of the ND-IRIS-0405 iris image dataset. This dataset is a superset of the iris image datasets used in ICE 2005 and ICE 2006.
  

• FRVT 2006 and ICE 2006 Large-Scale Experimental Results,
  P. Jonathon Phillips, W. Todd Scruggs, Alice O'Toole, Patrick J. Flynn, Kevin W. Bowyer, Cathy L. Schott and Matthew Sharpe,
  IEEE Transactions on Pattern Analysis and Machine Intelligence, to appear.
  pdf of this paper.
  DOI link.
  This paper describes the large-scale experimental results from the Face Recognition Vendor Test (FRVT) 2006 and the Iris Challenge Evaluation (ICE) 2006. ...
  

• Empirical Evidence for Correct Iris Match Score Degradation With Increased Time Lapse Between Gallery and Probe Images,
  Sarah Baker, Kevin W. Bowyer and Patrick J. Flynn,
  International Conference on Biometrics, June 2009, 1170-1179.
  pdf of this paper.
  We explore the effects of time lapse on iris biometrics using a data set of images with four years time lapse between the earliest and the most recent images of an iris (13 subjects, 26 irises, 1809 total images. We find that the average fractional distance for a match between two images of an iris taken four years apart is significantly larger than the match for images with only a few months time lapse between them. ... To our knowledge, this is the first and only experimental study of iris match scores under long (multi-year) time lapse.
  

• Overview of the Multiple Biometric Grand Challenge,
  P. Jonathon Phillips, Todd Scruggs, Patrick Flynn, Kevin W. Bowyer, Ross Beveridge, Geoff Givens, Bruce Draper and Alice O'Toole,
  International Conference on Biometrics, June 2009, 705-714.
  pdf of this paper.
  The goal of the Multiple Biometric Grand Challenge (MBGC) is to improve the performance of face and iris recognition technology from samples acquired under unconstrained conditions. The MBGC is organized into three challenge problems. Each challenge problem relaxes the constraints in different directions. ...
  

• Image Averaging for Improved Iris Recognition,
  Karen Hollingsworth, Kevin W. Bowyer and Patrick J. Flynn,
  International Conference on Biometrics, June 2009, 1112-1121.
  pdf of this paper.
  We take advantage of the temporal continuity in an iris video to improve matching performance using signal-level fusion. From multiple frames of an iris video, we create a single average image. Our signal-level fusion method performs better that methods based on single still images, and better than previously published multi-gallery score-fusion methods. ...
  

• Recent Research Results In Iris Biometrics,
  Karen Hollingsworth, Sarah Baker, Sarah Ring, Kevin W. Bowyer and Patrick J. Flynn,
  SPIE 7306B: Biometric Technology for Human Identification VI, April 2009.
  pdf of this paper.
  ... we have collected more than 100,000 iris images for use in iris biometrics research. Using this data, we have developed a number of techniques for improving recognition rates. These techniques include fragile bit masking, signal-level fusion of iris images, and detecting local distortions in iris texture. Additionally, we have shown that large degrees of dilation and long lapses of time between image acquisitions negatively impact performance.
  

• Introduction to the Special Section of Best Papers from the 2007 Biometrics: Theory, Applications and Systems Conference,
  Kevin W. Bowyer,
  IEEE Transactions on Systems, Man and Cybernetics - Part A, 39 (1), January 2009, 2-3.
  pdf of this paper.
  DOI link.
  ... Over 100 papers were submitted to BTAS 07. ... The final result of this process is the set of five papers that appear in this special section. We are particularly fortunate in the way that the five papers in this special section illustrate the breadth of activities in current biometrics research. Face, fingerprint, iris, voice, handwriting, and multimodal biometrics are all represented. ...
  

• Pupil Dilation Degrades Iris Biometric Performance,
  Karen Hollingsworth, Kevin W. Bowyer and Patrick J. Flynn,
  Computer Vision and Image Understanding, 113 (1), January 2009, 150-157.
  pdf of this paper.
  DOI link.
  ... We found that when the degree of dilation is similar at enrollment and recognition, comparisons involving highly dilated pupils result in worse recognition performance than comparisons involving constricted pupils. We also found that when the matched images have similarly highly dilated pupils, the mean Hamming distance of the match distribution increases and the mean Hamming distance of the non-match distribution decreases, bringing the distributions closer together from both directions. We further found that when matching enrollment and recognition images of the same person, larger differences in pupil dilation yield higher template dissimilarities, and so a greater chance of a false non-match. ...
  Our research group is part of the Multiple Biometric Grand Challenge and Iris Challenge Evaluation support teams.
  

• The Best Bits in an Iris Code,
  Karen Hollingsworth, Kevin W. Bowyer and Patrick J. Flynn,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 31 (6), 964-973, June 2009.
  pdf of this paper.
  ... The fractional Hamming distance weights all bits in an iris code equally. However, not all the bits in an iris code are equally useful. Our research is the first to present experiments documenting that some bits are more consistent than others. ... The inconsistencies are largely due to the coarse quantization of the phase response. Masking iris code bits corresponding to complex filter responses near the axes of the complex plane improves the separation between the match and nonmatch Hamming distance distributions.
  Our research group is part of the Multiple Biometric Grand Challenge and Iris Challenge Evaluation support teams.
  

• Using Multi-Instance Enrollment to Improve Performance 3D Face Recognition,
  Timothy C. Faltemier, Kevin W. Bowyer and Patrick J. Flynn,
  Computer Vision and Image Understanding 112 (2), November 2008, 114-125.
  Preprint pdf version of this paper.
  DOI link.
  This paper explores the use of multi-instance enrollment as a means to improve the performance of 3D face recognition. Experiments are performed using the ND-2006 3D face data set which contains 13,450 scans of 888 subjects. This is the largest 3D face data set currently available and contains a substantial amount of varied facial expression. Results indicate that the multi-instance enrollment outperforms a state-of-the-art component-based recognition approach ...
  

• Detection of Iris Texture Distortions By Analyzing Iris Code Matching Results,
  Sarah Ring and Kevin W. Bowyer,
  Biometrics: Theory, Applications and Systems (BTAS 08), September 2008, Washington, DC.
  pdf of this paper.
  Previous work in iris biometrics attempts to cope with occlusion by eyelids / eyelashes and with specular highlights through improved segmentation of the iris region. Our approach assumes that some local distortions of the iris texture are not detected at the segmentation stage, and that these generate corresponding regions of local distortion in the iris code derived from the image. We introduce an approach to detect such regions of local distortion in the iris code through analysis of the iris code matching results. We know of no previous work that attempts to detect distortions of iris texture through analyzing the iris code matching results.
  

• Multi-factor Approach To Improving Recognition Performance In Surveillance-quality Video,
  Deborah Thomas, Kevin W. Bowyer and Patrick J. Flynn,
  Biometrics: Theory, Applications and Systems (BTAS 08), September 2008, Washington, DC.
  DOI link.
  

• Profile Face Detection: A Subset Multi-biometric Approach,
  James Gentile, Kevin W. Bowyer and Patrick J. Flynn,
  Biometrics: Theory, Applications and Systems (BTAS 08), September 2008, Washington, DC.
  DOI link.
  

• The Iris Challenge Evaluation 2005,
  P. Jonathon Phillips, Kevin W. Bowyer and Patrick J. Flynn, Xiaomei Liu and W. Todd Scruggs,
  Biometrics: Theory, Applications and Systems (BTAS 08), September 2008, Washington, DC.
  DOI link.
  

• A Region Ensemble for 3D Face Recognition,
  Timothy Faltemier, Kevin W. Bowyer and Patrick J. Flynn,
  IEEE Transactions on Information Forensics and Security, 3(1):62-73, March 2008.
  DOI link.
  ... we introduce a new system for 3D face recognition based on the fusion of results from a committee of regions that have been independently matched. ... Rank-one recognition rates of 97.2% and verification rates of 93.2% at 0.1% false accept rate are reported and compared to other methods published on the Face Recognition Grand Challenge v2 data set.
  

• Image Understanding for Iris Biometrics: A Survey,
  Kevin W. Bowyer, Karen Hollingsworth and Patrick J. Flynn,
  Computer Vision and Image Understanding, 110(2), 281-307, May 2008.
  DOI link.
  ... Most research publications can be categorized as making their primary contribution to one of the four major modules in iris biometrics: image acquisition, iris segmentation, texture analysis and matching of texture representations. Other important research includes experimental evaluations, image databases, applications and systems, and medical conditions that may affect the iris. ...
  Our research group is part of the Multiple Biometric Grand Challenge and Iris Challenge Evaluation support teams.
  

• Using Classifier Ensembles to Label Spatially Disjoint Data,
  Larry Shoemaker, Robert E. Banfield, Lawrence O. Hall, Kevin W. Bowyer and W. Philip Kegelmeyer,
  Information Fusion 9(1), 120-133, January 2008.
  pdf of this paper.
  We describe an ensemble approach to learning from arbitrarily partitioned data. ... We combine a fast ensemble learning algorithm with probabilistic majority voting in order to learn an accurate classifier from such data. ...
  

• Learning to Predict Gender from Irises,
  Vince Thomas, Nitesh V. Chawla, Kevin W. Bowyer and Patrick J. Flynn,
  IEEE International Conference on Biometrics: Theory, Applications, and Systems (BTAS 07), September 2007.
  pdf of this paper.
  This paper employs machine learning techniques to develop models that predict gender based on the iris texture features. ...
  

• Guest Editorial: Introduction to the Special Issue on Recent Advances in Biometric Systems,
  Kevin W. Bowyer, Venu Govindaraju and Nalini Ratha,
  IEEE Transactions on Systems, Man and Cybernetics - B 37 (5), October 2007.
  pdf of this paper.
  We are pleased to present 14 papers in this special issue devoted to recent advances in biometric systems. ...
  

• Comment on the CASIA version 1.0 Iris Dataset,
  P. Jonathon Phillips, Kevin W. Bowyer and Patrick J. Flynn,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 29 (10), October 2007.
  pdf of this paper.
  We note that the images in the CASIA version 1.0 iris dataset have been edited so that the pupil area is replaced by a circular region of uniform intensity. We recommend that this dataset is no longer used in iris biometrics research ...
  

• FRVT 2006 and ICE 2006 Large-Scale Results,
  P. J. Phillips, W. T. Scruggs, A. J. O'Toole, P. J. Flynn, K.W. Bowyer, C. L. Schott, and M. Sharpe.
  National Institute of Standards and Technology, NISTIR 7408, http://face.nist.gov, 2007.
  pdf of this report.
  This report describes the large-scale experimental results from the Face Recognition Vendor Test (FRVT) 2006 and the Iris Challenge Evaluation (ICE) 2006. ...
  Our research group is part of the Multiple Biometric Grand Challenge and Iris Challenge Evaluation support teams.
  

• A Fast Algorithm for ICP-based 3D Shape Biometrics,
  Ping Yan and Kevin W. Bowyer,
  Computer Vision and Image Understanding, 107 (3), 195-202, September 2007.
  pdf of this paper.
  ... we present a novel approach, called "Pre-computed Voxel Nearest Neighbor," to reduce the computational time for shape matching in a biometrics context. The approach shifts the heavy computation burden to the enrollment stage, which is done offline. Experiments in 3D ear biometrics with 369 subjects and 3D face biometrics with 219 subjects demonstrate the effectiveness of our approach.
  

• Biometric Recognition Using Three-dimensional Ear Shape,
  Ping Yan and Kevin W. Bowyer.
  IEEE Transactions on Pattern Analysis and Machine Intelligence 29 (8), 1297-1308, August 2007.
  pdf of this paper.
  ... We present a complete system for ear biometrics, including automated segmentation of the ear in a profile view image and 3D shape matching for recognition. We evaluated this system with the largest experimental study to date in ear biometrics, achieving a rank-one recognition rate of 97.8% for an identification scenario, and equal error rate of 1.2% for a verification scenario on a database of 415 subjects and 1,386 total probes.
  

• Actively Exploring Face Space(s) for Improved Face Recognition,
  Nitesh V. Chawla and Kevin W. Bowyer,
  AAAI 2007, Vancouver, July 2007.
  pdf of this paper.
  We propose a learning framework that actively explores creation of face space(s) by selecting images that are complementary to the images already represented in the face space. We also construct ensembles of classifiers learned from such actively sampled image sets, which further provides improvement in the recognition rates. ...
  

• Boosting Lite - Handling Larger Datasets and Slower Base Classifiers,
  Lawrence O. Hall, Robert E. Banfield, Kevin W. Bowyer and W. Philip Kegelmeyer.
  Multiple Classifier Systems (MCS) 2007, Prague, May 2007.
  pdf of this paper.
  ... we examine ensemble algorithms (Boosting Lite and Ivoting) that provide accuracy approximating a single classifier, but which require significantly fewer training examples. ...
  

• A Comparison of Decision Tree Ensemble Creation Techniques,
  Robert E. Banfield, Lawrence O. Hall, Kevin W. Bowyer, and W. Philip Kegelmeyer.
  IEEE Transactions on Pattern Analysis and Machine Intelligence 29 (1), 173-180, January 2007.
  pdf of this paper. appendix to the paper.
  We experimentally evaluate bagging and seven other randomization-based approaches to creating an ensemble of decision tree classifiers. Statistical tests were performed on experimental results from 57 publicly available data sets. ...
  

• Face Recognition Using 2D, 3D and Infra-Red: Is Multi-modal Better than Multi-sample?
  Kevin W. Bowyer, Kyong I. Chang, Patrick J. Flynn and Xin Chen.
  Proceedings of the IEEE, 94 (11), 2000-2012, November 2006.
  pdf of this paper.
  We compare the performance improvement obtained by combining three-dimensional or infra-red with normal intensity images (a multi-modal approach) to the performance improvement obtained by using multiple intensity images (a multi-sample approach). Combining results from different types of imagery gives significantly higher recognition rates than are obtained by using a single intensity image. However, significantly higher recognition rates are also obtained by combining results from multiple intensity images.
  

• Multiple Nose Region Matching for 3D Face Recognition Under Varying facial expression,
  Kyong I. Chang, Kevin W. Bowyer, and Patrick J. Flynn,
  IEEE Transactions on Pattern Analysis and Machine Intelligence, 28 (10), 1695-1700, October 2006.
  pdf of this paper.
  An algorithm is proposed for 3D face recognition in the presence of varied facial expressions. It is based on combining the match scores from matching multiple overlapping regions around the nose. Experimental results are presented using the largest database employed to date in 3D face recognition studies, over 4,000 scans of 449 subjects. ...
  

• Multi-modal Biometrics: An Overview,
  Kevin W. Bowyer, et al,
  Second Workshop on Multi-Modal User Authentication (MMUA 2006), May 2006, Toulouse, France.
  pdf of this paper.
  The topic of multi-modal biometrics has attracted strong interest in recent years. This paper categorizes approaches to multi-modal biometrics based on the biometric source, the type of sensing used, and the depth of collaborative interaction in the processing. This paper also attempts to identify some of the challenges and issues that confront research in multi-modal biometrics.
  This paper represents the invited talk given to open the first day of the workshop.
  

• A Survey of Approaches and Challenges in 3D and Multi-modal 3D+2D Face Recognition,
  Kevin W. Bowyer, Kyong Chang, and Patrick J. Flynn,
  Computer Vision and Image Understanding 101 (1), January 2006, 1-15.
  pdf of this paper.
  ... This survey focuses on face recognition performed by matching models of the three-dimensional shape of the face, either alone or in combination with matching corresponding two-dimensional intensity images.
  This article was number one on the CVIU "Top 25" list for the quarters October - December 2005 and January - March of 2006, and in the top ten for seven consecutive quarters.
  

• Experiments With an Improved Iris Segmentation Algorithm,
  Xiaomei Liu, Kevin W. Bowyer, and Patrick J. Flynn,
  Fourth IEEE Workshop on Automatic Identification Advanced Technologies (AutoID), October 2005, New York, 118-123.
  pdf of this paper.
  ... We have also developed and implemented an improved iris segmentation and eyelid detection stage of the algorithm, and experimentally verified the improvement in recognition performance using the collected dataset. Compared to Masek's original segmentation approach, our improved segmentation algorithm leads to an increase of over 6% in the rank-one recognition rate.
  

• Infra-Red and Visible-Light Face Recognition,
  Xin Chen, Patrick J. Flynn, and Kevin W. Bowyer,
  Computer Vision and Image Understanding 99 (3), September 2005, 332-358.
  pdf of this paper.
  ... We find that in a scenario involving time lapse between gallery and probe, and relatively controlled lighting, (1) PCA-based recognition using visible images outperforms PCA-based recognition using infra-red images, and (2) the combination of PCA-based recognition using visible and infra-red imagery substantially outperforms either one individually...
  This article was number two on the CVIU "Top 25" list for July - September of 2005.
  

• Ensembles of Classifiers from Spatially Disjoint Data,
  Robert E. Banfield, Lawrence O. Hall, Kevin W. Bowyer, and W. Philip Kegelmeyer,
  Springer-Verlag LNCS 3541: 6th International Workshop on Multiple Classifier Systems (MCS 2005), Monterey, CA, June 2005, 196-205.
  pdf of this paper.
  ... We describe an ensemble learning approach that accurately learns from data which has been partitioned according to the arbitrary spatial requirements of a large-scale simulation wherein classifiers may be trained only the data local to a given partition. As a result, the class statistics can vary from partition to partition; some classes may even be missing from some partitions.
  

• Empirical Evaluation of Advanced Ear Biometrics,
  Ping Yan and Kevin W. Bowyer,
  Workshop on Empirical Evaluation Methods in Computer Vision, San Diego, CA, June 2005.
  pdf of this paper.
  We present results of the largest experimental investigation of ear biometrics to date. Approaches considered include a PCA ("eigen-ear") approach with 2D intensity images, achieving 63.8% rank-one recognition; a PCA approach with range images, achieving 55.3% Hausdorff matching of edge images from range images, achieving 67.5% and ICP matching of the 3D data, achieving 98.7%. ICP based matching not only achieves the best performance, but also shows good scalability with size of dataset. The data set used represents over 300 persons, each with images acquired on at least two different dates. In addition, the ICP-based approach is further applied on an expanded data set of 404 subjects, and achieves 97.5% rank one recognition rate. In order to test the robustness and variability of ear biometrics, ear symmetry is also investigated. In our experiments around 90% of peoples' right and left ears are symmetric.
  

• Overview of the Face Recognition Grand Challenge,
  P. Jonathon Phillips, Patrick J. Flynn, Todd Scruggs, Kevin W. Bowyer, Jin Chang, Kevin Hoffman, Joe Marques, Jaesik Min, and William Worek,
  Computer Vision and Pattern Recognition (CVPR 2005), San Diego, June 2005, I:947-954.
  pdf of this paper.
  Our research group is part of the Face Recognition Grand Challenge support team.
  

• Random Subspaces and Subsampling for 2-D Face Recognition,
  Nitesh V. Chawla and Kevin W. Bowyer,
  Computer Vision and Pattern Recognition (CVPR 2005) , San Diego, June 2005, II: 582-589.
  pdf of this paper.
  Random subspaces are a popular ensemble construction technique that improves the accuracy of weak classifiers. It has been shown, in different domains, that random subspaces combined with weak classifiers such as decision trees and nearest neighbor classifiers can provide an improvement in accuracy. In this paper, we apply the random subspace methodology to the 2-D face recognition task. The main goal of the paper is to see if the random subspace methodology can do as well, if not better, than the single classifier constructed on the tuned face space. We also propose the use of a validation set for tuning the face space, to avoid bias in the accuracy estimation. In addition, we also compare the random subspace methodology to an ensemble of subsamples of image data. This work shows that a random subspaces ensemble can outperform a well-tuned single classifier for a typical 2-D face recognition problem. The random subspaces approach has the added advantage of requiring less careful tweaking.
  

• An Evaluation of Multi-modal 2D+3D Face Biometrics,
  Kyong I. Chang, Kevin W. Bowyer, and Patrick J. Flynn,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 27 (4), April 2005, 619-624.
  pdf of this paper.
  We report on the largest experimental study to date in multi-modal 2D+3D face recognition ... Major conclusions are: (1) 2D and 3D have similar recognition performance when considered individually, (2) Combining 2D and 3D results using a simple weighting scheme outperforms either 2D or 3D alone, (3) Combining results from two or more 2D images using a similar weighting scheme also outperforms a single 2D image, and (4) Combined 2D+3D outperforms the multi-image 2D result. This is the first (so far, only) work to present such an experimental control to substantiate multi-modal performance improvement."
  

• Ensemble Diversity Measures and Their Application to Thinning,
  Robert E. Banfield, Lawrence O. Hall, Kevin W. Bowyer, and W. Philip Kegelmeyer,
  Information Fusion 6 (1), March 2005, 49-62.
  pdf of this paper.
  ... We evaluate thinning algorithms on ensembles created by several techniques on 22 publicly available datasets. When compared to other methods, our percentage correct diversity measure algorithm shows a greater correlation between the increase in voted ensemble accuracy and the diversity value. ... Finally, the methods proposed for thinning again show that ensembles can be made smaller without loss in accuracy.
  

• Improved Range Image Segmentation by Analyzing Surface Fit Patterns,
  Jaesik Min and Kevin W. Bowyer,
  Computer Vision and Image Understanding 97(2), February 2005, 242-258.
  pdf of this paper.
  We propose a new approach to range image segmentation of planar and curved surface scenes. Our method is mainly an extended design of an existing algorithm, which was guided by a framework of performance evaluation. We choose the range segmentation algorithm developed by Jiang and Bunke as our baseline algorithm, which is fast and has shown relatively high performance in several experimental performance evaluation studies. We analyze the types of errors made by the algorithm, propose design modifications to decrease the error rate, and experimentally verify that the new approach achieves statistically significant performance improvement. Whereas the baseline algorithm applies the edge-linking uniformly to all edge pixels to segment a region, the modified algorithm selects high potential edge areas in the region by analyzing the surface fit pattern and gives priority of edge-linking to those areas. The contributions of this work are (1) an improved algorithm for segmentation of range images of both planar and curved surface scenes, and (2) a demonstration of using empirical performance evaluation to guide algorithm design and modification to achieve better performance.
  

• The Human ID Gait Challenge Problem: Data Sets, Performance, and Analysis,
  Sudeep Sarkar, P. Jonathon Phillips, Zongyi Liu, Isidro Robledo, Patrick Grother, and Kevin W. Bowyer,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 27 (2), February 2005, 162-177.
  pdf of this paper.
  ... To provide a means for measuring progress and characterizing the properties of gait recognition, we introduce the HumanID Gait Challenge Problem. The challenge problem consists of a baseline algorithm, a set of 12 experiments, and a large data set. The baseline algorithm estimates silhouettes by background subtraction and performs recognition by temporal correlation of silhouettes. The 12 experiments are of increasing difficulty, as measured by the baseline algorithm, and examine the effects of five covariates on performance. ...
  

• Comments on "A Parallel Mixture of SVMs for Very Large Scale Problems,"
  Xiaomei Liu, Lawrence O. Hall, and Kevin W. Bowyer,
  Neural Computation 16 (7), July 2004, 1345-1351.
  pdf of this paper.
  ... Experiments on the Forest Cover data set show that this parallel mixture is more accurate than a single SVM, with 90.72% accuracy reported on an independent test set. While this accuracy is impressive, the referenced paper does not consider alternative types of classifiers. In this comment, we show that a simple ensemble of decision trees results in a higher accuracy, 94.75%, and is computationally efficient. This result is somewhat surprising and illustrates the general value of experimental comparisons using different types of classifiers.
  

• Learning Ensembles from Bites: a Scalable and Accurate Approach,
  Nitesh Chawla, Lawrence O. Hall, Kevin W. Bowyer and W. Philip Kegelmeyer,
  Journal of Machine Learning Research 5, April 2004, 421-451.
  pdf of this paper.
  ... Voting many classifiers built on small subsets of data is a promising approach for learning from massive data sets, one that can utilize the power of boosting and bagging. We propose a framework for building hundreds or thousands of such classifiers on small subsets of data in a distributed environment. Experiments show this approach is fast, accurate, and scalable.
  

• Face Recognition Technology and the Security Versus Privacy Tradeoff,
  Kevin W. Bowyer,
  IEEE Technology and Society, Spring 2004, 9-20.
  pdf of this paper.
  Video surveillance and face recognition systems have become the subject of increased interest and controversy after the September 11 terrorist attacks on the United States. ... This paper analyzes the interplay of technical and social issues involved in the widespread application of video surveillance for person identification.
  This paper received a 2005 Award of Excellence from the Society for Technical Communication.
  

• Automated Performance Evaluation of Range Image Segmentation Algorithms,
  Jaesik Min, Mark Powell, and Kevin W. Bowyer,
  IEEE Transactions on Systems, Man, and Cybernetics - Part B, 34 (1), February 2004, 263-271.
  pdf of this paper.
  ... We present an automated framework for evaluating the performance of range image segmentation algorithms. Automated tuning of algorithm parameters in this framework results in performance as good as that previously obtained with careful manual tuning by the algorithm developers. ... This framework is demonstrated using range images, but in principle it could be used to evaluate region segmentation algorithms for any type of images.
  

• Assessment of Time Dependency in Face Recognition: An Initial Study,
  Patrick Flynn, Kevin W. Bowyer and P. Jonathon Phillips,
  Audio- and Video-Based Biometric Person Authentication (AVBPA 2003), Springer Lecture Notes in Computer Science 2688, 44-51.
  pdf of this paper.
  ... Experimental results suggest that (a) recognition performance is substantially poorer when unknown images are acquired on a different day from the enrolled images, (b) degradation in performance does not follow a simple predictable pattern with time between known and unknown image acquisition, and (c) performance figures quoted in the literature based on known and unknown image sets acquired on the same day may have little practical value.
  

• Is Error-based Pruning Redeemable?,
  Lawrence O. Hall, Kevin W. Bowyer, Robert E. Banfield and Steven Eschrich, and Richard Collins,
  International Journal of Artificial Intelligence Tools, 12 (3), September 2003, 249-264.
  pdf of this paper.
  

• Comparison and Combination of Ear and Face Images for Appearance-based Biometrics,
  Kyong Chang, Kevin W. Bowyer, Sudeep Sarkar, and Barnabas Victor,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 25 (9), September 2003, 1160-1165.
  pdf of this paper.
  In the experiments reported here, recognition performance is essentially identical using ear images or face images and combining the two for multimodal recognition results in a statistically significant performance improvement. ... To our knowledge, ours is the only work to present any experimental results of computer algorithms for biometric recognition based on the ear.
  

• Face Recognition Using 2D and 3D Facial Data,
  Kyong I. Chang, Kevin W. Bowyer and Patrick J. Flynn,
  First Workshop on Multi-Modal User Authentication , Santa Barbara, 25-32, December 2003.
  pdf of this paper.
  Results are presented for the largest experimental study to date that investigates the comparison and combination of 2D and 3D face recognition. To our knowledge, this is the only such study to incorporate significant time lapse between gallery and probe image acquisition ...
  

• Distributed Learning with Bagging-like Performance,
  Nitesh Chawla, Thomas E. Moore, Lawrence O. Hall, Kevin W. Bowyer, W. Philip Kegelmeyer, and Clayton Springer,
  Pattern Recognition Letters 24 (1-3), 2003, 455-471.
  pdf of this paper.
  Bagging forms a committee of classifiers by bootstrap aggregation of training sets from a pool of training data. A simple alternative to bagging is to partition the data into disjoint subsets. Experiments with decision tree and neural network classifiers on various datasets show that, given the same size partitions and bags, disjoint partitions result in performance equivalent to, or better than, bootstrap aggregates (bags). Many applications (e.g., protein structure prediction) involve use of datasets that are too large to handle in the memory of the typical computer. Hence, bagging with samples the size of the data is impractical. Our results indicate that, in such applications, the simple approach of creating a committee of n classifiers from disjoint partitions each of size 1=n (which will be memory resident during learning) in a distributed way results in a classifier which has a bagging-like performance gain. The use of distributed disjoint partitions in learning is significantly less complex and faster than bagging.
  

• SMOTE: Synthetic Minority Over-sampling TEchnique,
  Nitesh Chawla, Kevin W. Bowyer, Lawrence O. Hall, and W. Philip Kegelmeyer,
  Journal of Artificial Intelligence Research 16, 2002, 321-357.
  pdf of this paper.
  This paper shows that a combination of our method of over-sampling the minority (abnormal) class and under-sampling the majority (normal) class can achieve better classifier performance (in ROC space) than only under-sampling the majority class. This paper also shows that a combination of our method of over-sampling the minority class and under-sampling the majority class can achieve better classifier performance (in ROC space) than varying the loss ratios in Ripper or class priors in Naive Bayes. Our method of over-sampling the minority class involves creating synthetic minority class examples.
  

• The Gait Identification Challenge Problem: Data Sets and Baseline Algorithm,
  P. Jonathon Phillips, Sudeep Sarkar, Isidro Robledo, Patrick Grother, and Kevin W. Bowyer,
  International Conference on Pattern Recognition (ICPR 2002) , August 2002, Montreal, I:385-388.
  

• ``Star Wars'' Revisited - A Continuing Case Study In Ethics and Safety-Critical Software,
  Kevin W. Bowyer,
  IEEE Technology and Society 21 (1), Spring 2002, 13-26.
  pdf of this paper.
  The Reagan-era Strategic Defense Initiative was the focus of a great deal of technical argument relating to the design and testing of safety critical software. ... This paper describes a curriculum module developed around a Reagan-era SDI debate on the theme - 'Star Wars: Can the computing requirements be met?' This module should be appropriate for use in ethics-related or software-engineering-related courses taught in undergraduate Information Systems, Information Technology, Computer Science, or Computer Engineering programs.
  

• Edge Detector Evaluation Using Empirical ROC Curves,
  Kevin W. Bowyer, Christine Kranenburg, and Sean Dougherty.
  Computer Vision and Image Understanding 84 (1), October 2001, 77-103.
  pdf of this paper.
  This paper focuses on evaluating the performance of edge detection algorithms. ... Performance is summarized using receiver operating characteristic curves. ... No other work uses adaptive parameter sampling to construct empirical ROC curves for pixel-level performance evaluation, or has used so many real images, or has compared such a broad selection of detectors.
  

• Comparison of Edge Detector Performance Through Use In An Object Recognition Task,
  Min C. Shin, Dmitry B. Goldgof and Kevin W. Bowyer.
  Computer Vision and Image Understanding 84 (1), October 2001, 160-178.
  DOI link.
  Edge detector performance is measured using a particular edge-based object recognition algorithm as a higher-level task. A detector's performance is ranked according to the object recognition performance that it generates. We have used a challenging train and test dataset containing 110 images of jeep-like objects. Six edge detectors are compared and results suggest that (1) the SUSAN edge detector performs best and (2) the ranking of various edge detectors is different from that found in other evaluations.
  

• Comparison of Edge Detection Algorithms Using a Structure From Motion Task,
  Min C. Shin, Dmitry B. Goldgof, Kevin W. Bowyer and S. Nikiforou.
  IEEE Transactions on Systems, Man and Cybernetics - Part B 31 (4), August 2001, 589-601.
  DOI link.
  We use the task of structure from motion as a "black box" through which to evaluate the performance of edge detection algorithms. Edge detector goodness is measured by how accurately the SFM could recover the known structure and motion from the edge detection of the image sequences. We use a variety of real image sequences with ground truth to evaluate eight different edge detectors from the literature. Our results suggest that ratings of edge detector performance based on pixel-level metrics and on the SFM are well correlated and that detectors such as the Canny detector and the Heitger detector offer the best performance.
  

• Mentoring undergraduates in computer vision research,
  Mubarak Shah and Kevin W. Bowyer,
  IEEE Transactions on Education 44 (3), 252-257, August 2001.
  DOI link.
  ... During the last 14 years roughly 130 undergraduate students from several institutions have participated in the research experiences for undergraduates (REU) program funded by the National Science Foundation (NSF). A large fraction of our students have been able to prepare a paper for submission to a conference, have the paper accepted, and then attend the conference to present the paper. ...
  

• Ethics and Computing: Living Responsibly In a Computerized World,
  Kevin W. Bowyer.
  IEEE Press (second edition), 2001.
  

• The Digital Database for Screening Mammography,
  Michael Heath, Kevin Bowyer, Daniel Kopans, Richard Moore and W. Philip Kegelmeyer.
  in Proceedings of the Fifth International Workshop on Digital Mammography, M.J. Yaffe, ed., 212-218, Medical Physics Publishing, 2001.
  pdf of this paper.
  The Digital Database for Screening Mammography (DDSM) is a database of digitized film-screen mammograms with associated ground truth information. The purpose of this resource is to provide a large set of mammograms in a digital format that may be used by researchers to evaluate and compare the performance of computer-aided detection (CAD) algorithms. ...
  

• Resources For Teaching Ethics and Computing,
  Kevin W. Bowyer,
  Journal of Information Systems Education 11 (3-4), 91-92, Summer-Fall 2000.
  

• Pornography On the Dean's PC: An Ethics and Computing Case Study,
  Kevin W. Bowyer,
  Journal of Information Systems Education 11 (3-4), 121-126, Summer-Fall 2000.
  JISE link.
  

• Themes for Improved Teaching of Image Computation,
  Kevin W. Bowyer, George Stockman and Louise Stark,
  IEEE Transactions on Education 43 (2), 221-223, May 2000.
  DOI link.
  This paper reports on recommendations and results from a panel discussion and a workshop devoted to the theme of education in areas that involve image computation. One specific set of contributions is in the area of integrating image computation into required core courses such as Introduction to Computing and Data Structures. Another area of specific contributions is the development of undergraduate elective courses on topics such as robotics and medical image analysis. A third area of contributions is the improvement of traditional image processing and computer vision courses.
  

• Registration and Difference Analysis of Corresponding Nammogram Images,
  Maha Y. Sallam and Kevin W. Bowyer,
  Medical Image Analysis 3 (2), 103-118, 1999.
  DOI link.
  An automated technique is proposed for identifying differences between corresponding mammogram images. The technique recovers an approximate deformation between a pair of mammograms based on identifying corresponding features across the two images. The registration process is completed using an unwarping technique for transforming one image into the coordinate system of the other. A difference image is generated using intensity-weighted subtraction in order to identify regions of large difference. Evaluation of the technique is performed using 124 bilateral image pairs which contain a total of 77 abnormalities of different types. The purpose of this paper is to measure the extent to which the mammogram registration technique is able to provide useful information for identifying abnormalities in mammograms.
  

• Evaluation of Texture Segmentation Algorithms,
  Kyong Chang, Kevin W. Bowyer and S. Munishkumaran,
  Computer Vision and Pattern Recognition (CVPR '99), I:294-299.
  pdf of this paper.
  This paper presents a method of evaluating unsupervised texture segmentation algorithms. The control scheme of texture segmentation has been conceptualized as two modular processes: (1) feature computation and (2) segmentation of homogeneous regions based on the feature values. Three feature extraction methods are considered: gray level co-occurrence matrix, Laws' texture energy and Gabor multi-channel filtering. Three segmentation algorithms are considered: fuzzy c-means clustering, square-error clustering and split-and-merge. A set of 35 real scene images with manually-specified ground truth was compiled. Performance is measured against ground truth on real images using region-based and pixel-based performance metrics.
  

• Dynamic-Scale Model Construction From Range Imagery,
  Adam Hoover Dmitry Goldgof, and Kevin W. Bowyer,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 29 (12), 1352-1357, December 1998.
  DOI link.
  The construction of a surface model from range data may be undertaken at any point in a continuum of scales that reflects the level of detail of the resulting model. This continuum relates the construction parameters to the scale of the model. We propose methods to dynamically reprocess range data at different scales. The construction result from a single scale is automatically evaluated, causing reconstruction at a different scale when user-defined criteria are not met. We demonstrate our methods in constructing a planar b-rep space envelope (a scene representation) for over 400 range images. The experiments demonstrate that ability to construct 100 percent valid models, with the scale of detail within specified requirements.
  

• Overview of Work In Empirical Evaluation of Computer Vision Algorithms,
  Kevin W. Bowyer and P. Jonathon Phillips,
  in Empirical Evaluation Techniques in Computer Vision, K.W. Bowyer and P.J. Phillips, editors, IEEE Computer Society Press, 1998, pages 1-11.
  pdf version of this chapter.
  

• Function from visual analysis and physical interaction: a methodology for recognition of generic classes of objects,
  Melanie Sutton, Louise Stark and Kevin Bowyer, Image and Vision Computing 16 (11), 745-763, August 1998.
  DOI link.
  This paper presents an overview of the GRUFF-I (Generic Recognition Using Form, Function and Interaction) system, a nonpart-based approach to generic object recognition which reasons about and generates plans for interaction with three-dimensional (3D) shapes from the categories furniture and dishes. ...
  

• Current Status of the Digital Database for Screening Mammography,
  Michael Heath, Kevin Bowyer, Daniel Kopans, W. Philip Kegelmeyer, Richard Moore, Kyong Chang and S. Munishkumaran,
  in Digital Mammography, 457-460, Kluwer Academic Publishers, 1998 (proceedings of the Fourth International Workshop on Digital Mammography)
  pdf of this paper.
  The Digital Database for Screening Mammography is a resource for use by researchers investigating mammogram image analysis. In particular, the resource is focused on the context of image analysis to aid is screening for breast cancer. The database now contains substantial numbers of "normal" and "cancer" cases. ...
  

• Are Edges Sufficient for Object Recognition?,
  Thomas A. Sanocki, Kevin W. Bowyer, Michael Heath, and Sudeep Sarkar.
  Journal of Experimental Psychology: Human Perception and Performance 24 (1), 340-349, January 1998.
  pdf of this paper.
  The authors argue that the concept of `edges' as used in current research on object recognition obscures the significant difficulties involved in interpreting stimulus information. ... With 1-s exposures, the accuracy of identifying objects in the edge images was found to be less than half that with color photographs. Therefore, edges are far from being sufficient for object recognition.
  

• Comparison of Edge Detectors: A Methodology and Initial Study,
  Michael Heath, Sudeep Sarkar, Thomas A. Sanocki and Kevin W. Bowyer.
  Computer Vision and Image Understanding 69, (1), 38-54, January 1998.
  DOI link.
  Because of the difficulty of obtaining ground truth for real images, the traditional technique for comparing low-level vision algorithms is to present image results, side by side, and to let the reader subjectively judge the quality. This is not a scientifically satisfactory strategy. However, human rating experiments can be done in a more rigorous manner to provide useful quantitative conclusions. We present a paradigm based on experimental psychology and statistics, in which humans rate the output of low level vision algorithms. We demonstrate the proposed experimental strategy by comparing four well-known edge detectors: ...
  

• A Robust Visual Method for Assessing the Relative Performance of Edge-Detection Algorithms,
  Michael Heath, Sudeep Sarkar, Thomas A. Sanocki, and Kevin W. Bowyer.
  IEEE Transactions on Pattern Analysis and Machine Intelligence 19 (12), 1338-1359, December 1997.
  pdf of this paper.
  A new method for evaluating edge detection is presented ... The basic measure of performance is a visual rating score which indicates the perceived quality of the edges for identifying an object. ... The novel aspect of this work is the use of a visual task and real images of complex scenes in evaluating edge detectors.
  

• Generating ROC curves for artificial neural networks,
  Kevin S. Woods and Kevin W. Bowyer,
  IEEE Transactions on Medical Imaging 16 (3), 329-337, June 1997.
  DOI link.
  Receiver operating characteristic (ROC) analysis is an established method of measuring diagnostic performance in medical imaging studies. Traditionally, artificial neural networks (ANN's) have been applied as a classifier to find one "best" detection rate. Recently researchers have begun to report ROC curve results for ANN classifiers. The current standard method of generating ROC curves for an ANN is to vary the output node threshold for classification. Here, the authors propose a different technique for generating ROC curves for a two class ANN classifier. They show that this new technique generates better ROC curves in the sense of having greater area under the ROC curve (AUC), and in the sense of being composed of a better distribution of operating points.
  

• Combination of Multiple Classifiers Using Local Accuracy Estimates,
  Kevin S. Woods, W. Philip Kegelmeyer and Kevin W. Bowyer,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 19 (4), 405-410, April 1997.
  pdf of this paper.
  This paper presents a method for combining classifiers that uses estimates of each individual classifier's local accuracy in small regions of the feature space surrounding an unknown sample. An empirical evaluation using five real data sets confirms the validity of our approach compared to some other combination of multiple classifiers algorithms. We also suggest a methodology for determining the best mix of individual classifiers.
  

• Recognizing object function through reasoning about partial shape descriptions and dynamic physical properties,
  Louise Stark, Kevin W. Bowyer, Adam W. Hoover and Dmitry B. Goldgof,
  Proceedings of the IEEE 84 (11), 1640-1656, November 1996.
  DOI link.
  Knowledge about required functionality of an object can be used as an effective representation for a generic object category (e.g., "chair", "cup", or "hammer"). This approach to object representation and recognition has recently become an active area of research. We explore a scenario in which a robot senses the environment to obtain an initial partial shape model of an object. If the information in this initial model is not sufficient to hypothesize a possible function for the object, then additional view(s) may be suggested. Once a possible function is hypothesized, a plan is formulated for interacting with the object to confirm that its material properties are compatible with the hypothesized function. The module for reasoning about partial shape models has been evaluated on over 200 shape models acquired from range images. The module for carrying out a function verification plan has been evaluated in a simulated environment using the ThingWorld (TW) system.
  

• An Experimental Comparison of Range Image Segmentation Algorithms,
  Adam W. Hoover, Gillian Jean-Baptiste, Xiaoyi Jiang, Patrick Flynn, Horst Bunke, Dmitry Goldgof, Kevin W. Bowyer, David Eggert, Andrew Fitzgibbon, and Robert Fisher,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 18, (7), 673-689, July 1996.
  pdf of this paper.
  This paper evaluates four segmentation algorithms on 80 real images with ground truth and objective performance measures. ... This type of framework for a competitive effort is essentially never used in mainstream computer vision, though it is standard practice in some related areas ... Beside the development of a philosophy of comparative experiment research, an important contribution here is an assessment of the state-of-the-art in planar range image segmentation. Based on our results, we assert that this problem is not 'solved.' This finding may be surprising and possibly controversial. We would welcome an empirical demonstration that the claim is false.
  

• Learning Membership Functions in a Function-based Object Recognition System,
  Kevin S. Woods, Diane Cook, Lawrence Hall, Louise Stark and Kevin W. Bowyer,
  Journal of Artificial Intelligence Research 3, 187-222, October 1995.
  

• On recovering hyperquadrics from range data,
  Senthil Kumar, Song Han, Dmitry Goldgof, and Kevin W. Bowyer,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 17, (11), 1079-1083, November 1995. Computer Vision and Image Understanding 62 (2), 177-193, September 1995.
  

• Extracting a Valid Boundary Representation From a Segmented Range Image,
  Adam W. Hoover, Dmitry Goldgof, and Kevin W. Bowyer,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 17, (9), 920-924, September 1995.
  A new approach is presented for extracting an explicit 3-D shape model from a single range image. One novel aspect is that the model represents both observed object surfaces, and surfaces which bound the volume of occluded space. Another novel aspect is that the approach does not require that the range image segmentation be perfect. ... A third novel aspect of this work is that the implementation has been evaluated on over 200 real images of polyhedral objects with no operator intervention and all parameters held constant, and obtained a 97% success rate in creating valid b-reps.
  DOI link.
  

• Generic Recognition of Articulated Objects Through Reasoning About Potential Function,
  Kevin Green, David Eggert, Louise Stark, and Kevin W. Bowyer,
  Computer Vision and Image Understanding 62 (2), 177-193, September 1995.
  

• Aspect Graphs and Their Use in Object Recognition,
  David Eggert, Louise Stark, and Kevin W. Bowyer,
  Annals of Mathematics and Artificial Intelligence 13, 347-375, 1995.
  

• GRUFF-3: Generalizing the Domain of a Function-based Recognition System,
  Melanie Sutton, Louise Stark, and Kevin W. Bowyer, Pattern Recognition 27 (12), 1743-1766, December 1994.
  

• Function-based Generic Recognition for Multiple Object Categories,
  Louise Stark and Kevin W. Bowyer,
  CVGIP: Image Understanding 59 (1), 1-21, January 1994.
  ... Our work concentrates specifically on the relation between shape and function of rigid 3D objects. Recognition of an observed shape is performed by reasoning about the function that it might serve. Previous efforts have dealt with only a single basic level object category. A number of important issues arise in extending this approach to deal with multiple basic-level categories. ...
  

• Comparative evaluation of pattern recognition techniques for detection of microcalcifications in mammography,
  Kevin S. Woods, Christopher C. Doss, Kevin W. Bowyer, Jeffrey L. Solka, Carey E. Priebe, and W. Philip Kegelmeyer,
  International Journal of Pattern Recognition and Artificial Intelligence 7 (6), 1417-1436, December 1993.
  Computer-assisted detection of microcalcifications in mammographic images will likely require a multistage algorithm that includes segmentation of possible microcalcifications, pattern recognition techniques to classify the segmented objects, a method to determine if a cluster of calcifications exists, and possibly a method to determine the probability of malignancy. This paper focuses on the first three of these stages ...
  

• The Scale Space Aspect Graph,
  David W. Eggert, Kevin W. Bowyer, Charles R. Dyer, Henrik I. Christensen, and Dmitry B. Goldgof.
  IEEE Transactions on Pattern Analysis and Machine Intelligence 15 (11), 1114-1130, November 1993.
  pdf of this paper.
  This paper introduces the concept of the scale space aspect graph, defines three different interpretations of the scale dimension, and presents a detailed example for a simple class of objects, with scale defined in terms of the spatial extent of features in the image.
  

• An Investigation of Methods of Combining Functional Evidence for 3-D Object Recognition,
  Louise Stark, Lawrence O. Hall, and Kevin W. Bowyer,
  International Journal of Pattern Recognition and Artificial Intelligence 7 (3), 573-594, June 1993.
  

• Computing the Generalized Aspect Graph for Objects with Moving Parts,
  Kevin W. Bowyer, Maha Y. Sallam, David Eggert, and John H. Stewman,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 15 (6), 605-610, June 1993.
  A number of researchers have described algorithms for computing the aspect graph representation, but this work has thus far been limited to entirely rigid objects. We generalize this concept to include a larger, more realistic domain of objects known as articulated assemblies, those objects composed of rigid parts with articulated connections allowed between parts.
  

• Computing the Perspective Projection Aspect Graph of Solids of Revolution,
  David W. Eggert and Kevin W. Bowyer,
  IEEE Transactions on Pattern Analysis and Machine Intelligence 15 (2), 109-128, February 1993.
  This paper presents the first (only) implemented algorithm to compute the aspect graph for a class of curved-surface objects based on an exact parcellation of 3-D viewpoint space. The class of objects considered is solids of revolution. ... the worst-case complexity of ... the number of cells in the aspect graph is is O(N^4), where N is the degree of a polynomial that defines the object shape.
  

• Achieving Generalized Object Recognition Through Reasoning About Association of Function To Structure,
  Louise Stark and Kevin W. Bowyer.
  IEEE Transactions on Pattern Analysis and Machine Intelligence 13 (10), 1097-1104, October 1991.
  pdf of this paper.
  The purpose of the work described here is to demonstrate the feasibility of defining an object category in terms of the functional properties shared by all objects in the category. ... A complete system has been implemented that takes the boundary surface description of a 3-D object as its input and attempts to recognize whether the object belongs to the category chair ... This is, to our knowledge the first (only) implemented system to explore the use of a purely function-based definition of an object category ... to recognize 3-D objects.
  

• Aspect Graphs: An Introduction and Survey of Recent Results,
  Kevin W. Bowyer and Charles R. Dyer,
  International Journal of Imaging Systems and Technologies 2, 315-328, 1990.
  

• Direct Construction of Perspective Projection Aspect Graphs for Planar-Face Convex Objects,
  John Stewman and Kevin W. Bowyer,
  Computer Vision, Graphics and Image Processing 51 (1), 20-37, July 1990.
  

• The Effects of Age on Radionuclide Angiographic Detection and Quantitation of Left-to-right Shunts,
  Page A.W. Anderson, Kevin W. Bowyer, and Robert H. Jones,
  American Journal of Cardiology 53, 879-883, March 1984.
  

• Optimizing Contiguous Element Region Selection for Virtual Memory Computer Systems,
  Kevin W. Bowyer, and C. Frank Starmer,`
  IEEE Transactions on Computers 32, 1201-1203, December 1983.
  

• Radionuclide Analysis of Right and Left Ventricular Response to Exercise in Patients with Atrial and Ventricular Septal Defects,
  Claude A. Peter, Kevin W. Bowyer, and Robert H. Jones,
  American Heart Journal 105, 428-435, March 1983.
  

• Error Sensitivity of Computed Tomography Guided Stereotaxis,
  Kevin W. Bowyer, C. Frank Starmer and P. J. DuBois,
  Computers and Biomedical Research 15, 272-280, 1982.
  

• A Simulation-based Sensitivity Study of Angiocardiographic Approaches to Shunt Assessment,
  Kevin W. Bowyer and C. Frank Starmer,
  Computers and Biomedical Research 15, 111-128, 1982.
  

• CT-guided Stereotaxis Using a Modified Conventional Stereotaxic Frame,
  P. J. DuBois, B. S. Nashold, J. Perry, P. Burger, Kevin W. Bowyer, E. R. Heinz, B. P. Drayer, S. Bigner, and A. C. Higgins,
  American Journal of Neuroradiology 3, 345-351, 1982.