Wavelet Feature Maps Compression for Image-to-Image CNNs
This addresses the computational resource challenge for CNNs in image-to-image tasks, offering a more efficient compression method with reduced performance loss, though it is incremental as it builds on existing quantization and wavelet transform techniques.
The paper tackles the problem of severe performance degradation in image-to-image tasks like semantic segmentation and depth estimation when using aggressive quantization (less than 4-bits) for compressing CNNs, and proposes Wavelet Compressed Convolution (WCC) to compress high-resolution activation maps, achieving compression rates equivalent to 1-4bit activation quantization with relatively small and graceful degradation in performance.
Convolutional Neural Networks (CNNs) are known for requiring extensive computational resources, and quantization is among the best and most common methods for compressing them. While aggressive quantization (i.e., less than 4-bits) performs well for classification, it may cause severe performance degradation in image-to-image tasks such as semantic segmentation and depth estimation. In this paper, we propose Wavelet Compressed Convolution (WCC) -- a novel approach for high-resolution activation maps compression integrated with point-wise convolutions, which are the main computational cost of modern architectures. To this end, we use an efficient and hardware-friendly Haar-wavelet transform, known for its effectiveness in image compression, and define the convolution on the compressed activation map. We experiment with various tasks that benefit from high-resolution input. By combining WCC with light quantization, we achieve compression rates equivalent to 1-4bit activation quantization with relatively small and much more graceful degradation in performance. Our code is available at https://github.com/BGUCompSci/WaveletCompressedConvolution.