https://github.com/AaltoVision/Object-Retrieval

Particular object retrieval using CNN
https://github.com/AaltoVision/Object-Retrieval

Last synced: 22 days ago
JSON representation

Particular object retrieval using CNN

• Host: GitHub
• URL: https://github.com/AaltoVision/Object-Retrieval
• Owner: AaltoVision
• License: mit
• Created: 2017-03-03T15:45:12.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2017-03-10T14:07:20.000Z (over 7 years ago)
• Last Synced: 2024-05-14T00:10:52.559Z (about 1 month ago)
• Language: Python
• Size: 8.79 KB
• Stars: 17
• Watchers: 6
• Forks: 9
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Lists

• awesome-cbir-papers - Particular object retrieval using CNN
• awesome-image-retrieval-papers - Particular object retrieval using CNN

README

        
# Object-Retrieval

Particular object retrieval using CNN

Code for the arxiv submission https://arxiv.org/pdf/1703.01226.pdf. 

The source code is a modification of the codes available at http://www.xrce.xerox.com/Our-Research/Computer-Vision/Learning-Visual-Representations/Deep-Image-Retrieval. Follow the instructions in the above link to setup the installations. Then replace the file "test.py" with "SA.py", and the same set of commands provided in their README should work fine. However, to reproduce the results in our paper, the command associated with multi-resolution should be used (explained in detail in README from the Xerox-link).