Architecture-aware Network Pruning for Vision Quality Applications
This work addresses efficiency issues for vision quality applications like low-light imaging and super-resolution, though it appears incremental as it builds on existing pruning methods.
The paper tackles the high computational complexity of CNNs in vision quality applications by proposing an architecture-aware pruning algorithm, achieving 58% and 37% reductions in MAC for low-light imaging and super-resolution without quality drop.
Convolutional neural network (CNN) delivers impressive achievements in computer vision and machine learning field. However, CNN incurs high computational complexity, especially for vision quality applications because of large image resolution. In this paper, we propose an iterative architecture-aware pruning algorithm with adaptive magnitude threshold while cooperating with quality-metric measurement simultaneously. We show the performance improvement applied on vision quality applications and provide comprehensive analysis with flexible pruning configuration. With the proposed method, the Multiply-Accumulate (MAC) of state-of-the-art low-light imaging (SID) and super-resolution (EDSR) are reduced by 58% and 37% without quality drop, respectively. The memory bandwidth (BW) requirements of convolutional layer can be also reduced by 20% to 40%.