Skip to content

Pytorch implementation of pedestrian attribute recognition with graph convolutional network

Notifications You must be signed in to change notification settings

2014gaokao/pedestrian-attribute-recognition-with-GCN

Repository files navigation

pedestrian-attribute-recognition-with-GCN

Preparation

Prerequisite: Python 3.6 and torch 1.1.0 and tqdm

Download RAP(v2) dataset and annotation then put in dataset directory

Train the model

( If you simply want to run the demo code without further modification, you might skip this step by downloading the weight file from Baidu Yun with password "5z1j" and put the model_best.pth.tar into directory /checkpoint/ then run
python demo.py )

python transform_rap2.py     (transform data)
python glove.py      (word2vec)
python adj.py      (Adjacency matrix)
python train.py      (weight file will locate in checkpoint directory)

Methodology

image

Superiority

method mA accuracy precision recall F1
ACN 69.66 62.61 80.12 72.26 75.98
DeepMar 73.79 62.02 74.92 76.21 75.56
HP-Net 76.12 65.39 77.33 78.79 78.05
JRL 77.81 - 78.11 78.98 78.58
VeSPa 77.70 67.35 79.51 79.67 79.59
Ours 75.97 68.99 81.48 79.97 80.72

About

Pytorch implementation of pedestrian attribute recognition with graph convolutional network

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages