refinenet

We have hosted the application refinenet in order to run this application in our online workstations with Wine or directly.

Quick description about refinenet:

RefineNet is a MATLAB-based framework for semantic image segmentation and general dense prediction tasks. It implements the architecture presented in the CVPR 2017 paper RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation and its extended version published in TPAMI 2019. The framework uses multi-path refinement and improved residual pooling to achieve high-quality segmentation results across multiple benchmark datasets. It provides trained models for datasets such as PASCAL VOC 2012, Cityscapes, NYUDv2, Person_Parts, PASCAL_Context, SUNRGBD, and ADE20k, with versions based on ResNet-101 and ResNet-152 backbones. The repository supports both single-scale and multi-scale prediction, with scripts for training, testing, and evaluating segmentation performance. While this codebase is specific to MATLAB and MatConvNet, a PyTorch implementation and lighter-weight variants are also available from the community.

Features:

• Implements RefineNet for high-resolution semantic segmentation
• Provides trained models on seven benchmark datasets
• Supports single-scale and multi-scale prediction with fusion
• Uses improved residual pooling for better segmentation accuracy
• Includes training and evaluation scripts for custom datasets
• Compatible with ResNet-101 and ResNet-152 backbones in MatConvNet

Programming Language: C++, MATLAB, Python, Unix Shell.
Categories: