Fast image compression with neural networks beats JPEG in speed and quality
PIC: Revisiting INR for Image Coding with Fast Encoding and Sub-Millisecond Decoding
Computer Vision and Pattern Recognition
Summary
Images need to be compressed so they take up less space but still look good when viewed. Traditional methods like JPEG are fast to decode but sometimes not the best quality, while newer neural network methods can give better quality but take too long to encode and decode. The authors developed a new neural approach called PIC that speeds up both encoding and decoding, allowing images to be compressed nearly as fast as JPEG with similar or better quality. Their system can encode images quickly and then decode them extremely fast, making it practical for real use. They also made their code available for others to try.
Implicit neural representation (INR)Image compressionEncoding speedDecoding speedRate-distortion performanceJPEGNeural networksImage codecsFeedforward architecture
Authors
Xiang Liu, Jinxiang Wang, Bin Chen, Zimo Liu, Mingyao Hong, Jiawei Li, Yaowei Wang, Shu-tao Xia
Abstract
Implicit neural representation (INR) has achieved remarkable progress in novel view synthesis and image/video coding in recent years.Compared to conventional end-to-end image codecs, INR-based compressors demonstrate significant advantages in decoding complexity. However, their practical application has been hindered by the inferior encoding speed and underutilized decoding efficiency.In this work, we propose a feedforward INR image coding architecture, Practical INR Image Codec (PIC), that computes all the necessary information for INR network in a single forward pass, achieving an encoding speed of 20 FPS. Additionally, we implement a highly optimized decoder that reaches 2000 FPS decoding speed, significantly surpassing JPEG's performance at comparable rate-distortion (RD) performance. To the best of our knowledge, this work presents the first learning-based image codec that simultaneously outperforms or is comparable with JPEG in both RD performance and decoding speed while maintaining practical encoding speed. Code is available at https://github.com/actcwlf/PIC.