LWRN: Light-Weight Residual Network for Edge Detection

Edge detection is one of the most fundamental fields in computer vision. With the rapid development of the combination of Convolutional Neural Network and Multi-Scale Representation of image, significant progress has been made in this field. However, most of them have a huge size, which makes it hard to apply in reality, and a huge number of parameters may lead to waste of computing resources. In this paper, we focus on qualitative analysis of the role of each part in the network, and propose a modified light-weight architecture based on our result and the study of former works. Our new architecture is composed of residual-blocks, max-pooling layers and batch normalization layers. Compared with the previous models, the new architecture performs better in memory, convergence and computation efficiency with similar model size. Moreover, the new architecture can achieve better accuracy with smaller model size. When evaluating our model on the well-known BSDS500 benchmark, we achieve ODS F-measure of 0.769 with parameters less than 0.3 M, which shows a better property than the state-of-the-art result 0.766 at this level.

Automated summary

Edge detection is a fundamental field in computer vision, and significant progress has been made with the combination of Convolutional Neural Network and Multi-Scale Representation of image. However, the large size and number of parameters of existing models make them difficult to apply in reality and result in resource wastage. In this paper, a modified light-weight architecture is proposed based on qualitative analysis of each part in the network and previous research. The new architecture, composed of residual-blocks, max-pooling layers, and batch normalization layers, outperforms previous models in memory, convergence, and computation efficiency while maintaining similar model size. It achieves better accuracy with smaller model size and surpasses the state-of-the-art results on the BSDS500 benchmark.