Digital architecture for real-time CNN-based face detection for video processing

Abstract : In this paper, we propose a hardware computing architecture for face detection that classifies an image as a face or non-face. The computing architecture is first designed, modeled and tested in MATLAB Simulink using Xilinx block set and was later tested using a Virtex-6 FPGA ML605 Evaluation Kit. The system uses learned filters which were previously extracted by training on a set of face and non-face patterns. The system is fully feature based and does not require any assumptions on specific image processing techniques. The proposed approach takes an input image as a whole and passes it through different modules that apply sub-algorithms based on image convolution and sub-sampling followed by a non-linear signal processor containing artificial neurons. The architecture takes the form of a deep convolutional neural network (CNN) which can classify if a search window inside a picture contains a human face or not.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01574551
Contributor : Christophe Garcia <>
Submitted on : Tuesday, August 15, 2017 - 3:23:28 PM
Last modification on : Thursday, November 21, 2019 - 2:29:22 AM

Identifiers

Citation

Smrity Bhattarai, Arjuna Madanayake, Renato J. Cintra, Stefan Duffner, Christophe Garcia. Digital architecture for real-time CNN-based face detection for video processing. IEEE Cognitive Communications for Aerospace Applications Workshop (CCAA), Jun 2017, Cleveland, United States. ⟨10.1109/CCAAW.2017.8001608⟩. ⟨hal-01574551⟩

Share

Metrics

Record views

235