chinefed/convolutional-set-transformer

Official implementation of the Convolutional Set Transformer (Chinello & Boracchi, 2025). This repository includes the source code of the cstmodels Python package, which provides reusable Keras 3 layers for constructing CST architectures, together with an interface to load and use the CST-15 model pre-trained on ImageNet.

22
/ 100
Experimental

This project offers a new way for machine learning engineers and researchers to process collections of related images, even if they look different and there are varying numbers of images in each collection. It takes in sets of images and can classify them or detect anomalies within the sets, providing explainable results. This is for users building advanced image analysis systems who need to interpret how their models make decisions.

No commits in the last 6 months.

Use this if you need to analyze collections of images, where each collection represents a single concept or category, and you want a model that performs better than existing methods while also being interpretable.

Not ideal if you are working with individual images rather than sets, or if you don't require explainable AI capabilities for your image analysis tasks.

image-set-classification anomaly-detection-images computer-vision deep-learning-research explainable-ai
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 5 / 25
Maturity 15 / 25
Community 0 / 25

How are scores calculated?

Stars

11

Forks

Language

Jupyter Notebook

License

MIT

Last pushed

Sep 30, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/chinefed/convolutional-set-transformer"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.