# Refining Image Categorization by Exploiting Web Images and General Corpus

## TL;DR

This paper introduces a method to improve image categorization by automatically selecting and classifying web images into semantic subcategories using general corpus data, addressing label noise and classifying beyond nouns.

## Contribution

It proposes a novel approach that leverages general corpus information and multi-instance learning to refine image categories without relying on labor-intensive labeled datasets.

## Key findings

- Significant performance improvements in image categorization and sub-categorization.
- Outperforms existing weakly supervised and web-supervised methods.
- Effective noise reduction in both subcategory labels and web images.

## Abstract

Studies show that refining real-world categories into semantic subcategories contributes to better image modeling and classification. Previous image sub-categorization work relying on labeled images and WordNet's hierarchy is not only labor-intensive, but also restricted to classify images into NOUN subcategories. To tackle these problems, in this work, we exploit general corpus information to automatically select and subsequently classify web images into semantic rich (sub-)categories. The following two major challenges are well studied: 1) noise in the labels of subcategories derived from the general corpus; 2) noise in the labels of images retrieved from the web. Specifically, we first obtain the semantic refinement subcategories from the text perspective and remove the noise by the relevance-based approach. To suppress the search error induced noisy images, we then formulate image selection and classifier learning as a multi-class multi-instance learning problem and propose to solve the employed problem by the cutting-plane algorithm. The experiments show significant performance gains by using the generated data of our way on both image categorization and sub-categorization tasks. The proposed approach also consistently outperforms existing weakly supervised and web-supervised approaches.