论文信息 - Multiple-clue face detection algorithm using edge-based feature vectors

Multiple-clue face detection algorithm using edge-based feature vectors

A multiple-clue image perception has been developed, aimed at direct implementation in a VLSI hardware accelerator, and was applied to the problem of face detection. In the algorithm, feature vectors are generated, utilizing the distribution of edges in an input image, thus achieving dimensionality reduction for efficient processing. In addition, multiple clues in the edge distribution are utilized to enhance the accuracy of face detection. For this purpose, several feature vector generation schemes have been developed, which are all compatible with the direct hardware implementation. In software simulation, over 91% of human faces have been detected correctly with only 20 face templates, and the number of false positives has been reduced drastically by the multiple-clue scheme.

Tadashi Shibata | Yasufumi Suzuki

[1] Takeo Kanade,et al. Neural Network-Based Face Detection , 1998, IEEE Trans. Pattern Anal. Mach. Intell..

[2] Chengjun Liu,et al. Gabor feature based classification using the enhanced fisher linear discriminant model for face recognition , 2002, IEEE Trans. Image Process..

[3] Tadashi Shibata,et al. A Mixed-Signal VLSI for Real-Time Generation of Edge-Based Image Vectors , 2003, NIPS.

[4] T. Shibata,et al. A general-purpose vector-quantization processor employing two-dimensional bit-propagating winner-take-all , 2002, 2002 Symposium on VLSI Circuits. Digest of Technical Papers (Cat. No.02CH37302).

[5] Tadashi Shibata,et al. Analog Soft-Pattern-Matching Classifier using Floating-Gate MOS Technology , 2001, NIPS.

[6] Tadashi Shibata,et al. An image representation algorithm compatible with neural-associative-processor-based hardware recognition systems , 2003, IEEE Trans. Neural Networks.