Contrastive Visual Data Augmentation

TL;DR

This paper introduces CoDA, a contrastive visual data augmentation method that enhances large multimodal models' ability to recognize novel and rare concepts by generating targeted synthetic data, significantly improving accuracy.

Contribution

The paper presents a novel contrastive data augmentation strategy that leverages multimodal generative models to improve recognition of unseen concepts in LMMs.

Findings

CoDA improves accuracy by up to 12.3% on NovelSpecies.

It outperforms existing data augmentation methods on multiple datasets.

Human verification confirms quality of augmented data.

Abstract

Large multimodal models (LMMs) often struggle to recognize novel concepts, as they rely on pre-trained knowledge and have limited ability to capture subtle visual details. Domain-specific knowledge gaps in training also make them prone to confusing visually similar, commonly misrepresented, or low-resource concepts. To help LMMs better align nuanced visual features with language, improving their ability to recognize and reason about novel or rare concepts, we propose a Contrastive visual Data Augmentation (CoDA) strategy. CoDA extracts key contrastive textual and visual features of target concepts against the known concepts they are misrecognized as, and then uses multimodal generative models to produce targeted synthetic data. Automatic filtering of extracted features and augmented images is implemented to guarantee their quality, as verified by human annotators. We show the effectiveness and efficiency of CoDA on low-resource concept and diverse scene recognition datasets including INaturalist and SUN. We additionally collect NovelSpecies, a benchmark dataset consisting of newly discovered animal species that are guaranteed to be unseen by LMMs. LLaVA-1.6 1-shot updating results on these three datasets show CoDA significantly improves SOTA visual data augmentation strategies by 12.3% (NovelSpecies), 5.1% (SUN), and 6.0% (iNat) absolute gains in accuracy.