JaveyWang / Pyramid-Attention-Networks-pytorch

Implementation of Pyramid Attention Networks for Semantic Segmentation.
GNU General Public License v3.0
236 stars 55 forks source link
attention pyramid semanticsegmentation

PAN-pytorch

A Pytorch implementation of Pyramid Attention Networks for Semantic Segmentation from 2018 paper by Hanchao Li, Pengfei Xiong, Jie An, Lingxue Wang. image

Installation

VOC2012 Dataset

The overall dataset is augmented by Semantic Boundaries Dataset, resulting in training data 10582 and test data 1449. https://www.sun11.me/blog/2018/how-to-use-10582-trainaug-images-on-DeeplabV3-code/

After preparing the data, please change the directory below for training.

training_data = Voc2012('/home/tom/DISK/DISK2/jian/PASCAL/VOC2012', 'train_aug', transform=train_transforms)
test_data = Voc2012('/home/tom/DISK/DISK2/jian/PASCAL/VOC2012', 'val',transform=test_transforms)

Evaluation

image

Pixel acc mIOU
93.19% 78.498%