6 papers
Leveraging Uncertainty for Deep Interpretable Classification and Weakly-Supervised Segmentation of Histology Images
Soufiane Belharbi, Jérôme Rony, Jose Dolz +3
Trained using only image class label, deep weakly supervised methods allow image classification and ROI segmentation for interpretability. Despite their success on natural images,…
F-CAM: Full Resolution Class Activation Maps via Guided Parametric Upscaling
Soufiane Belharbi, Aydin Sarraf, Marco Pedersoli +3
Class Activation Mapping (CAM) methods have recently gained much attention for weakly-supervised object localization (WSOL) tasks. They allow for CNN visualization and interpretati…
Deep Active Learning for Joint Classification & Segmentation with Weak Annotator
Soufiane Belharbi, Ismail Ben Ayed, Luke McCaffrey +1
CNN visualization and interpretation methods, like class-activation maps (CAMs), are typically used to highlight the image regions linked to class predictions. These models allow t…
Deep Interpretable Classification and Weakly-Supervised Segmentation of Histology Images via Max-Min Uncertainty
Soufiane Belharbi, Jérôme Rony, Jose Dolz +3
Weakly-supervised learning (WSL) has recently triggered substantial interest as it mitigates the lack of pixel-wise annotations. Given global image labels, WSL methods yield pixel-…
Non-parametric Uni-modality Constraints for Deep Ordinal Classification
Soufiane Belharbi, Ismail Ben Ayed, Luke McCaffrey +1
We propose a new constrained-optimization formulation for deep ordinal classification, in which uni-modality of the label distribution is enforced implicitly via a set of inequalit…
Min-max Entropy for Weakly Supervised Pointwise Localization
Soufiane Belharbi, Jérôme Rony, Jose Dolz +3
Pointwise localization allows more precise localization and accurate interpretability, compared to bounding box, in applications where objects are highly unstructured such as in me…