Efficient Super Resolution Using Binarized Neural Network

• Published in: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) pp. 694 - 703
• Main Authors: Ma, Yinglan, Xiong, Hongyu, Hu, Zhe, Ma, Lizhuang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2019
• Subjects: Image reconstruction; Image resolution; Laplace equations; Neural networks; Quantization (signal); Task analysis; Training
• ISSN: 2160-7516
• DOI: 10.1109/CVPRW.2019.00096

Deep convolutional neural networks (DCNNs) have recently demonstrated high-quality results in single-image super-resolution (SR). DCNNs often suffer from over-parametrization and large amounts of redundancy, which results in inefficient inference and high memory usage, preventing massive applications on mobile devices. As a way to significantly reduce model size and computation time, binarized neural network has only been shown to excel on semantic-level tasks such as image classification and recognition. However, little effort of network quantization has been spent on image enhancement tasks like SR, as network quantization is usually assumed to sacrifice pixel-level accuracy. In this work, we explore an network-binarization approach for SR tasks without sacrificing much reconstruction accuracy. To achieve this, we binarize the convolutional filters in only residual blocks, and adopt a learnable weight for each binary filter. We evaluate this idea on several state-of-the-art DCNN-based architectures, and show that binarized SR networks achieve comparable qualitative and quantitative results as their real-weight counterparts. Moreover, the proposed binarized strategy could help reduce model size by 80% when applying on SRResNet, and could potentially speed up inference by 5×.