Dedicated Inference Engine and Binary-Weight Neural Networks for Lightweight Instance Segmentation
This work addresses efficiency challenges for embedded systems by enabling lightweight instance segmentation with reduced hardware costs, though it is incremental as it builds on existing BNN and segmentation methods.
The authors tackled the problem of high computational costs in embedded systems by proposing a hardware architecture for binary-weight neural networks (BNNs) that reduces gate count and achieves 52% lower hardware costs compared to related works, while also developing lightweight instance segmentation networks that achieve higher accuracy than YOLACT with a 77.7× smaller model size.
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.