PiPS explains image segmentation models without retraining or loss
PiPS: Post-Hoc Prototypical Explanations for Interpretable Semantic Segmentation
Computer Vision and Pattern Recognition
Summary
Understanding why a computer program labels each pixel in a picture can be hard, especially when using complex neural networks. The authors present PiPS, a method that explains these decisions by showing parts of the image that look like prototypes, without needing to retrain or change the original model. This keeps the model's accuracy intact while making its reasoning easier to understand. PiPS works with any existing segmentation model, helping build safer and more trustworthy applications.
What this means in practice
- •For medical imaging teams: Provide clinicians with clear visual explanations for automated tissue or lesion segmentation without retraining the underlying models.
- •For autonomous vehicle developers: Generate trustworthy and spatially precise explanations for segmentation decisions in vehicle perception systems using existing pretrained networks.
Authors
Miłosz Adamczyk, Tymoteusz Zapala, Piotr Borycki, Przemysław Spurek
Abstract
With the increasing deployment of deep neural networks in critical systems, such as medical diagnostics and autonomous vehicles, ensuring their interpretability is crucial to building trust in decision-making systems. In the field of explainable artificial intelligence, prototype-based reasoning has gained particular popularity, as it mimics human cognitive processes by explaining model decisions based on visual similarity under the looks like this paradigm. While this paradigm has been thoroughly investigated in the context of global image classification, the interpretability of dense predictions, particularly semantic segmentation, remains largely unexplored despite its immense importance in tasks requiring precise object localization. Existing prototype-based interpretable segmentation models rely on ante-hoc architectures, which entails significant limitations because they require costly training from scratch and modifications to the network structure, ultimately leading to a noticeable drop in predictive performance compared to standard black-box models. To address this issue, we propose PiPS (Post-hoc interpretable Prototypical Segmentation), the first fully post-hoc solution for generating prototypical explanations for semantic segmentation models. Our method enables the extraction of intuitive, spatially localized explanations from any pre-trained network without modification or fine-tuning, thereby preserving 100% of the model's original predictive performance. This approach opens a new avenue for the safe and cost-effective deployment of transparent systems in advanced computer vision tasks. Codebase available at https://github.com/gmum/PIPS.