Dedicated Inference Engine and Binary-Weight Neural Networks for Lightweight Instance Segmentation

TL;DR

This paper introduces a specialized hardware inference engine for binary-weight neural networks that significantly reduces hardware costs and enables efficient lightweight instance segmentation with higher accuracy and much smaller model size.

Contribution

It proposes a novel hardware architecture for BNNs with two operation modes, reducing hardware costs by 52% and enabling practical lightweight instance segmentation networks.

Findings

Hardware costs reduced by 52% compared to related work.

Achieves higher accuracy than YOLACT on 'Person' category.

Model size is 77.7 times smaller than YOLACT.

Abstract

Reducing computational costs is an important issue for development of embedded systems. Binary-weight Neural Networks (BNNs), in which weights are binarized and activations are quantized, are employed to reduce computational costs of various kinds of applications. In this paper, a design methodology of hardware architecture for inference engines is proposed to handle modern BNNs with two operation modes. Multiply-Accumulate (MAC) operations can be simplified by replacing multiply operations with bitwise operations. The proposed method can effectively reduce the gate count of inference engines by removing a part of computational costs from the hardware system. The architecture of MAC operations can calculate the inference results of BNNs efficiently with only 52% of hardware costs compared with the related works. To show that the inference engine can handle practical applications, two lightweight networks which combine the backbones of SegNeXt and the decoder of SparseInst for instance segmentation are also proposed. The output results of the lightweight networks are computed using only bitwise operations and add operations. The proposed inference engine has lower hardware costs than related works. The experimental results show that the proposed inference engine can handle the proposed instance-segmentation networks and achieves higher accuracy than YOLACT on the "Person" category although the model size is 77.7$\times$ smaller compared with YOLACT.