serre-lab/Harmonization
👋 Aligning Human & Machine Vision using explainability
This project helps researchers and practitioners in AI and cognitive science compare how deep neural networks 'see' objects with how humans do. It takes images and a pre-trained neural network, then outputs an analysis of where the network focuses attention compared to human eye-tracking data. This is for AI researchers, cognitive scientists, and anyone developing or evaluating computer vision models who needs to understand and improve their human-likeness.
No commits in the last 6 months.
Use this if you need to understand, measure, and improve how closely your deep neural network's object recognition strategies align with human visual perception.
Not ideal if you are looking for a general-purpose image classification tool or a solution for non-vision AI tasks.
Stars
53
Forks
12
Language
Python
License
MIT
Last pushed
Jul 14, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/serre-lab/Harmonization"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
obss/sahi
Framework agnostic sliced/tiled inference + interactive ui + error analysis plots
tensorflow/tcav
Code for the TCAV ML interpretability project
MAIF/shapash
🔅 Shapash: User-friendly Explainability and Interpretability to Develop Reliable and Transparent...
TeamHG-Memex/eli5
A library for debugging/inspecting machine learning classifiers and explaining their predictions
csinva/imodels
Interpretable ML package 🔍 for concise, transparent, and accurate predictive modeling...