Bio-inspired hybrid framework for multi-view face detection

McCarroll, N; Belatreche, A; Harkin, J; Li, Y

Abstract

Reliable face detection in completely uncontrolled settings still remains a challenging task. This paper introduces a novel hybrid learning strategy that achieves robust in-plane and out-of-plane multi-view face detection through the enhanced implementation of the hierarchical bio-inspired HMAX framework using spiking neurons. Through multiple training trials, separate pools of neurons are trained on different face poses to extract features through feed-forward unsupervised STDP. The trained neurons are then processed by an additional STDP mechanism to generate a streamlined repository of broadly tuned multi-view neurons. After unsupervised feature extraction, supervised feature selection is implemented within the hybrid framework to reduce false positives. The hybrid system achieves robust invariant detection of in-plane and out-of-plane rotated faces that compares favourably with state-of-the-art face detection systems.

Citation

McCarroll, N., Belatreche, A., Harkin, J., & Li, Y. (2015). Bio-inspired hybrid framework for multi-view face detection. In Neural Information Processing : 22nd International Conference, ICONIP 2015, November 9-12, 2015, Proceedings, Part IV (232-239). Springer International Publishing. https://doi.org/10.1007/978-3-319-26561-2_28

• Publication Date: Nov 18, 2015
• Deposit Date: Jan 6, 2016
• Pages: 232-239
• Series Title: Lecture Notes in Computer Science
• Series Number: 9492
• Book Title: Neural Information Processing : 22nd International Conference, ICONIP 2015, November 9-12, 2015, Proceedings, Part IV
• ISBN: 9783319265605
• DOI: https://doi.org/10.1007/978-3-319-26561-2_28
• Publisher URL: http://dx.doi.org/10.1007/978-3-319-26561-2_28
• Related Public URLs: http://dx.doi.org/10.1007/978-3-319-26561-2