• Search by Module
• Search by Words
• Search Projects

• Java
• Python
• JavaScript
• TypeScript
• C++
• Scala
• Blog

Project: DBCNN-PyTorch (GitHub Link)

• DBCNN-PyTorch-master
  • SCNN3.py
  • NetVLAD.py
  • fc_models
    • keep.txt
  • BCNN.py
  • LIVEFolder.py
  • pretrained_scnn
    • pretrained_scnn.txt
  • DualCNN.py
  • WPFolder.py
  • dataset
    • dataset.txt
  • DBCNN.py
  • LICENSE
  • CBP.py
  • Koniq_10k.py
  • BaseCNN.py
  • LIVEChallengeFolder.py
  • TRILINEAR.py
  • README.md
  • MPNCOV.py
  • SCNN.py
  • .gitignore
  • simple_demo.py
  • loss
    • EMDLoss.py
    • FocalLoss.py
    • cross_entropy_prob.py
  • db_models
    • keep.txt

DBCNN-Pytorch

An experimental PyTorch implementation of Blind Image Quality Assessment Using A Deep Bilinear Convolutional Neural Network.

Purpose

Considering the popularity of PyTorch in academia, we hope this repo can help reseachers in IQA. This repo will be used as an active codebase for integrating advanced technologies for IQA research.

Requirements

PyTorch 0.4+ Python 3.6

Usage with default setting

python DBCNN.py

Only support experiment on LIVE IQA and LIVE Challenge right now, other datasets will be added soon! (I am a busy yet lazy guy...)

If you want to re-train the SCNN, you still need Matlab and original repo https://github.com/zwx8981/BIQA_Project for generating synthetically distorted images.

python SCNN.py

Acknowledgement

https://github.com/HaoMood/bilinear-cnn