Interpretable Image Classification with Adaptive Prototype-based Vision Transformers

TL;DR

ProtoViT introduces an interpretable image classification method that uses deformable prototypes within Vision Transformers, enabling better geometric matching, clearer explanations, and improved performance over existing prototype-based models.

Contribution

The paper integrates deformable prototypes with Vision Transformers for interpretable classification, offering adaptive parts and enhanced explanation fidelity.

Findings

Achieves higher accuracy than existing prototype models.

Prototypes are consistent and faithful in explanations.

Handles geometric variations effectively.

Abstract

We present ProtoViT, a method for interpretable image classification combining deep learning and case-based reasoning. This method classifies an image by comparing it to a set of learned prototypes, providing explanations of the form ``this looks like that.'' In our model, a prototype consists of \textit{parts}, which can deform over irregular geometries to create a better comparison between images. Unlike existing models that rely on Convolutional Neural Network (CNN) backbones and spatially rigid prototypes, our model integrates Vision Transformer (ViT) backbones into prototype based models, while offering spatially deformed prototypes that not only accommodate geometric variations of objects but also provide coherent and clear prototypical feature representations with an adaptive number of prototypical parts. Our experiments show that our model can generally achieve higher performance than the existing prototype based models. Our comprehensive analyses ensure that the prototypes are consistent and the interpretations are faithful.