Subjective quality assessment platform

• A web-based cloudsourcing platform using django-web

• Translated from Chinese to English by ron.lee

  • original : https://github.com/Max-cvv/IQA-website

• Characteristics

  • It shows two images and the user should choose one with better quality.
  • It allows zoom and panning of two images in a synchronised way.
  • It does not ask the user to score.
  • It is designed to compare the image quality of smartphone models, and the real time ranking of smartphone models can be found in the management backend. However, in the English translation, I changed the 'smartphone model' into '(IR) algorithm' to make it general.

Installation

• Django version should be below 3, or there would be errors.
• The python version the author tested was 3.6.9.

1. pip install -r requirements.txt

2. For the synchronized image comparasion, we use deepzoom Python Deep Zoom Tools. To install :
   cd deepzoom; python setup.py install; cd ..

Usage

• python manage.py runserver if you want to access locally
• python manage.py runserver 0.0.0.0:8000 if you want to access from another machine

Management Backend

• Add /manage/ to the URL. (id & passwd: admin).

• The provided functionalities are:

  • Dataset：view pictures. Add device. Upload pictures. When uploading pictures, name the pictures serially and we recommend jpg format 1.jpg, 2.jpg...), then zip them.
  • Experient info (User list, Record list, Question info, Real-time Ranking）：View subjective quality numbers
  • Experiment management

Pages

1. Main page

   

2. Experiment page

   

3. Management page