ROI-based Deep Image Compression with Swin Transformers
This addresses the problem of efficient image compression for applications like video conferencing and surveillance by focusing quality on key areas, though it is incremental as it builds on existing transformer methods.
The paper tackles image compression by encoding regions of interest (ROI) with higher quality than the background, using a Swin transformer-based autoencoder with a binary mask for spatial guidance and Lagrange multipliers to control region importance. It achieves higher ROI PSNR than other methods and shows superior object detection and instance segmentation performance on COCO.
Encoding the Region Of Interest (ROI) with better quality than the background has many applications including video conferencing systems, video surveillance and object-oriented vision tasks. In this paper, we propose a ROI-based image compression framework with Swin transformers as main building blocks for the autoencoder network. The binary ROI mask is integrated into different layers of the network to provide spatial information guidance. Based on the ROI mask, we can control the relative importance of the ROI and non-ROI by modifying the corresponding Lagrange multiplier $ λ$ for different regions. Experimental results show our model achieves higher ROI PSNR than other methods and modest average PSNR for human evaluation. When tested on models pre-trained with original images, it has superior object detection and instance segmentation performance on the COCO validation dataset.