Advancing Image Classification with Discrete Diffusion Classification Modeling
By: Omer Belhasin , Shelly Golan , Ran El-Yaniv and more
Potential Business Impact:
Helps computers guess pictures better, even when unsure.
Image classification is a well-studied task in computer vision, and yet it remains challenging under high-uncertainty conditions, such as when input images are corrupted or training data are limited. Conventional classification approaches typically train models to directly predict class labels from input images, but this might lead to suboptimal performance in such scenarios. To address this issue, we propose Discrete Diffusion Classification Modeling (DiDiCM), a novel framework that leverages a diffusion-based procedure to model the posterior distribution of class labels conditioned on the input image. DiDiCM supports diffusion-based predictions either on class probabilities or on discrete class labels, providing flexibility in computation and memory trade-offs. We conduct a comprehensive empirical study demonstrating the superior performance of DiDiCM over standard classifiers, showing that a few diffusion iterations achieve higher classification accuracy on the ImageNet dataset compared to baselines, with accuracy gains increasing as the task becomes more challenging. We release our code at https://github.com/omerb01/didicm .
Similar Papers
Learning Robust Diffusion Models from Imprecise Supervision
Machine Learning (CS)
Teaches computers to create better pictures from messy data.
Fitting Image Diffusion Models on Video Datasets
CV and Pattern Recognition
Makes AI create videos that look more real.
Class-N-Diff: Classification-Induced Diffusion Model Can Make Fair Skin Cancer Diagnosis
CV and Pattern Recognition
Makes AI create realistic skin disease pictures.