-
PyTorch 1.8.1 (with Python 3.7.0) implementation of the following paper. This version uses ResNet50 as the backbone network (Aesthetics Network), and you can replace it with other backbone network.
-
Demo.py for model test.
-
Pretrained weights: https://pan.baidu.com/s/1Tq73Io3PCTzQRGGzSwcJdA?pwd=t2vd
-
If you want to retrain the Visual Attribute Analysis Network (VAAN) and Theme Understanding Network (TUN), you need download AADB and EVA databases.
-
If you find our work is useful, pleaes cite our paper:
@ARTICLE{TAVAR,
author={Li, Leida and Huang, Yipo and Wu, Jinjian and Yang, Yuzhe and Li, Yaqian and Guo, Yandong and Shi, Guangming},
journal={IEEE Transactions on Circuits and Systems for Video Technology},
title={Theme-aware Visual Attribute Reasoning for Image Aesthetics Assessment},
year={2023},
doi={10.1109/TCSVT.2023.3249185}} -
If you have additional questions, you can email [email protected].
yipoh / tavar Goto Github PK
View Code? Open in Web Editor NEW[TCSVT] Theme-aware Visual Attribute Reasoning for Image Aesthetics Assessment