A Deep Image Compression Framework for Face Recognition

Abstract

Face recognition technology has advanced rapidly and has been widely used in various applications. Due to the extremely huge amount of data of face images and the large computing resources required correspondingly in large-scale face recognition tasks, there is a requirement for a face image compression approach that is highly suitable for face recognition tasks. In this paper, we propose a deep convolutional autoencoder compression network for face recognition tasks. In the compression process, deep features are extracted from the original image by the convolutional neural networks to produce a compact representation of the original image, which is then encoded and saved by existing codec such as PNG. This compact representation is utilized by the reconstruction network to generate a reconstructed image of the original one. In order to improve the face recognition accuracy when the compression framework is used in a face recognition system, we combine this compression framework with a existing face recognition network for joint optimization. We test the proposed scheme and find that after joint training, the Labeled Faces in the Wild (LFW) dataset compressed by our compression framework has higher face verification accuracy than that compressed by JPEG2000, and is much higher than that compressed by JPEG.

Compressing Face Images Using Genetic and Gray Wolf Meta-heuristic Algorithms Based on Variable Bit Allocation

“…The image of a face is one of the most popular and widely used images. There is a growing importance of the issue of identity verification in surveillance systems [20], control systems [21] and security systems [22]. Most of these systems require a large database of face images of different people, especially when they are used in large, international and significant organizations.…”

A Quick and Easy Way to Estimate Entropy and Mutual Information for Neuroscience

“…In short, it consists of two main steps: (i) bit reduction, replacing commonly used symbols with shorter representations and less commonly used symbols with longer representations by Huffman coding and (ii) Duplicate string elimination by detecting duplicates and replacing the occurrences by a reference to the first one, by LZSS algorithm. Its great efficacy made it a reference for comparison with other entropy-encoding image compression methods ( Cover and Thomas, 2006 ; Bian et al, 2019 ; Hou et al, 2020 ; Mentzer et al, 2020 ) and it is even used directly to estimate image entropy ( Wagstaff and Corsetti, 2010 ).…”

A quick and easy way to estimate entropy and mutual information for neuroscience

“…Its lossless compression is based on the combination of the Lempel–Ziv–Storer–Szymanski and Huffman algorithms and is called DEFLATE (Deutsch, 1996). Its great efficacy made it a reference for comparison with other entropy-encoding image compression methods (Bian et al, 2019; Cover and Thomas, 2006; Hou et al, 2020; Mentzer et al, 2020) and it is even used directly to estimate image entropy (Wagstaff and Corsetti, 2010).…”

“…In short, it consists of two main steps: i) bit reduction, replacing commonly used symbols with shorter representations and less commonly used symbols with longer representations by Huffman coding and ii) Duplicate string elimination by detecting duplicates and replacing the occurrences by a reference to the first one, by LZSS algorithm. Its great efficacy made it a reference for comparison with other entropy-encoding image compression methods (Bian et al, 2019;Cover and Thomas, 2006;Hou et al, 2020;Mentzer et al, 2020) and it is even used directly to estimate image entropy (Wagstaff and Corsetti, 2010).…”