https://github.com/manuelz/cards-identification

Cards identification system that matches a query card to a 106K card database using fine-tuned deep feature embeddings and similarity search.
https://github.com/manuelz/cards-identification

deep-learning faiss metric-learning opencv pairwise-similarity python pytorch

Last synced: 4 months ago
JSON representation

Cards identification system that matches a query card to a 106K card database using fine-tuned deep feature embeddings and similarity search.

• Host: GitHub
• URL: https://github.com/manuelz/cards-identification
• Owner: ManuelZ
• Created: 2024-09-04T18:29:37.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2025-05-29T01:18:37.000Z (4 months ago)
• Last Synced: 2025-06-14T23:53:38.598Z (4 months ago)
• Topics: deep-learning, faiss, metric-learning, opencv, pairwise-similarity, python, pytorch
• Language: Jupyter Notebook
• Homepage:
• Size: 34 MB
• Stars: 1
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Cards Identification

*Associated article: [Cards Identification project](https://manuelz.github.io/projects/cards-identification/)*

*Test the application: [Cards Identification Test App](https://huggingface.co/spaces/cestmanuel/cards-identification)*

Given a query card and a database of 106K cards, the goal is to determine which card in the database matches the query card.

![image](https://github.com/user-attachments/assets/f7821999-0585-45d2-888e-9f032496e599)

This experiment is inspired by the problem of face identification, which tries to answer the question: "Who is this person?"

[Take a look at the notebook](siamese_pt/Siamese_pytorch.ipynb).