Hideyasu Sai

Ear biometrics using jet space similarity

An ear biometric system that extends the ideas from elastic graph matching and principal component analysis (PCA) is introduced. For ear representation, an ldquoear graphrdquo whose vertices are labeled by the Gabor jets of body of the antihelix, superior antihelix crus, and inferior antihelix crus is introduced. A new ear detection algorithm is developed, introducing the notion of ldquojet space similarity,rdquo which denotes the similarity between Gabor jets and reconstructed jets obtained by PCA. Using the Extended Multi Modal Verification for Teleservices and Security applications (XM2VTS) database, the performance of the introduced system is found to be comparable to that of existing systems, and the detection performance is found to be higher compared with existing object detection algorithms using the distance from face space and average jets.

ICA, LDA, and Gabor Jets for Robust Ear Recognition, and Jet Space Similarity for Ear Detection

Abstract Fully automatic and robust ear recognition systems that use only 2D grayscale still images are presented. To exploit robustness against pose variation, changes in lighting, and hair occlusions, we thoroughly examined the techniques: linear discriminant analysis (LDA), independent component analysis (ICA), and Gabor jets. We obtained a 93.3% rank-one recognition rate on a dataset of 121 subjects in 4 image sets taken on various days from the public face database XM2VTS, where 47.0% of the images show hair occlusion, pose variation, and jagged images. To fully automate the recognition algorithm, we developed an ear detection algorithm that uses Gabor jets subjected to training using principal components analysis. A 1.0% equal error rate was obtained in experiments on the XM2VTS database. Our experiments provide evidence that ear biometrics has the potential for use in real-world applications for identifying individuals by their ears.

Improving the Robustness of Single-View Ear-Based Recognition under a Rotated in Depth Perspective

An algorithm is proposed for improving the robustness of an ear biometric system with the aim of developing a surveillance-system based on ear biometrics. To deal with pose variations that are rotated in depth, the Gabor jets of different poses are estimated and used as training data for a discriminant analysis-based classifier. Experimental evaluations show the effectiveness of the proposed algorithm, and the potential for improving the robustness of a single-view-based ear surveillance-system.

Estimating a Shooting Angle in Ear Recognition

To improve on our earlier work on single-view-based ear biometrics, an estimation method is presented for the shooting angle of an ear image based on the summation of similarity scores over a threshold within a database of known shooting angles. Experimental results indicate that the estimation method can improve the robustness of ear recognition in varied poses.

Improving ear recognition robustness from single-view-based images rotated in depth for forensic observations

Although ear shape is used in forensic investigations, ear identification system assisting forensic experts is not well developed. One of the reasons for this is ears 3D concave shape that changes their 2D appearances when there are changes in camera angles. To compensate these changes in 2D appearance, 3D statistical modeling is necessary. For this purpose, only one model is used in our previous study. Based on a few forensic classification methods of ear shapes, we attempted to increase the number of 3D statistical ear models. Through experimental evaluation, we find a possibility to improve the robustness against change in camera angles.