Lightweight high-resolution Subject Matting in the Real World
This work addresses the need for efficient and accurate subject matting in mobile and high-resolution scenarios, offering an incremental improvement over existing methods.
The paper tackles the problem of achieving fast and accurate high-resolution subject matting for real-world applications by proposing a lightweight network PSUNet and a new dataset HRSOM, resulting in superior performance over 13 existing methods with 113ms inference time on a mobile platform.
Existing saliency object detection (SOD) methods struggle to satisfy fast inference and accurate results simultaneously in high resolution scenes. They are limited by the quality of public datasets and efficient network modules for high-resolution images. To alleviate these issues, we propose to construct a saliency object matting dataset HRSOM and a lightweight network PSUNet. Considering efficient inference of mobile depolyment framework, we design a symmetric pixel shuffle module and a lightweight module TRSU. Compared to 13 SOD methods, the proposed PSUNet has the best objective performance on the high-resolution benchmark dataset. Evaluation results of objective assessment are superior compared to U$^2$Net that has 10 times of parameter amount of our network. On Snapdragon 8 Gen 2 Mobile Platform, inference a single 640$\times$640 image only takes 113ms. And on the subjective assessment, evaluation results are better than the industry benchmark IOS16 (Lift subject from background).