hackerschen / DSNet

DsNet: A Novel Hybrid Architecture of Convolution and Transformer for Real-time Weld Seam Image Segmentation

DSNet: A Dynamic Squeeze Network for Real-time Weld Seam Image Segmentation

This repository contains the official implementation of DSNet.

DSNet achieves superior trade-off between performance and latency

We present DSNet, an mannully designed semantic segmentation network with not only state-of-the-art performance but also faster speed than current methods.

The overall architecture of DSNet

Here is our speed-accuracy comparison on the weld seam dataset

Prerequisites

• Pytorch 2.1
• Python 3.8
• NVIDIA GPU

Installation

• Clone this repo:

git clone https://github.com/hackerschen/DSNet.git
cd DSNet

• Install dependencies:

pip install -r requirements.txt