CASSOD-Net: Cascaded and Separable Structures of Dilated Convolution for Embedded Vision Systems and Applications

TL;DR

This paper introduces CASSOD-Net, a new dilated convolution structure optimized for embedded vision systems, achieving higher accuracy and faster computation in face detection and image segmentation tasks.

Contribution

It proposes a cascaded and separable dilated convolution module and a specialized hardware system, improving efficiency and accuracy over traditional methods.

Findings

CASSOD-Net reduces filter weights by 53% in face detection.

Hardware acceleration is 2.78 times faster for dilated convolutions.

CASSOD maintains accuracy while improving efficiency.

Abstract

The field of view (FOV) of convolutional neural networks is highly related to the accuracy of inference. Dilated convolutions are known as an effective solution to the problems which require large FOVs. However, for general-purpose hardware or dedicated hardware, it usually takes extra time to handle dilated convolutions compared with standard convolutions. In this paper, we propose a network module, Cascaded and Separable Structure of Dilated (CASSOD) Convolution, and a special hardware system to handle the CASSOD networks efficiently. A CASSOD-Net includes multiple cascaded $2 \times 2$ dilated filters, which can be used to replace the traditional $3 \times 3$ dilated filters without decreasing the accuracy of inference. Two example applications, face detection and image segmentation, are tested with dilated convolutions and the proposed CASSOD modules. The new network for face detection achieves higher accuracy than the previous work with only 47% of filter weights in the dilated convolution layers of the context module. Moreover, the proposed hardware system can accelerate the computations of dilated convolutions, and it is 2.78 times faster than traditional hardware systems when the filter size is $3 \times 3$.