Abstract
Face recognition is one research area that has benefited from the recent popularity of deep learning, namely the convolutional neural network (CNN) model. Nevertheless, the recognition performance is still compromised by the model’s dependency on the scale of input images and the limited number of feature maps in each layer of the network. To circumvent these issues, we propose PSI-CNN, a generic pyramid-based scale-invariant CNN architecture which additionally extracts untrained feature maps across multiple image resolutions, thereby allowing the network to learn scale-independent information and improving the recognition performance on low resolution images. Experimental results on the LFW dataset and our own CCTV database show PSI-CNN consistently outperforming the widely-adopted VGG face model in terms of face matching accuracy.
Highlights
IntroductionDeepFace [1], DeepID [2], FaceNet [3], and the VGG model [4]
With the recent advancement in deep learning, a variety of models based on the convolutional neural network (CNN, see Figure 1) have been introduced in face recognition, most notablyDeepFace [1], DeepID [2], FaceNet [3], and the VGG model [4]
We evaluated the recognition performance of our proposed PSI-CNN model on two datasets—Labeled Faces in the Wild (LFW) [13] and our custom dataset derived from CCTV cameras
Summary
DeepFace [1], DeepID [2], FaceNet [3], and the VGG model [4] These models apply convolution and pooling repeatedly to extract local and global features of input image. The number of feature maps outputted by the shallower-side convolutional layers are limited by available memory and computational power, inevitably using only a subset of useful patterns for face recognition. This can be interpreted as excluding some potentially useful patterns, which we define as untrained patterns in this paper, thereby yielding suboptimal matching performance. Devising a way to reintroduce these untrained patterns into the model is the main motivation of this work
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
