DiVE-k: Differential Visual Reasoning for Fine-grained Image Recognition
This work improves fine-grained image recognition for applications requiring differentiation between visually similar categories, though it is incremental as it builds on existing reinforcement learning frameworks.
The paper tackles the problem of fine-grained image recognition by addressing the brittleness and memorization issues in existing reinforcement learning methods, proposing DiVE-k, which uses the model's top-k predictions to create multiple-choice questions for training, resulting in significant performance gains, such as surpassing QWEN2.5-VL-7B and ViRFT by 10.04% and 6.16% on the Harmonic Mean metric.
Large Vision Language Models (LVLMs) possess extensive text knowledge but struggles to utilize this knowledge for fine-grained image recognition, often failing to differentiate between visually similar categories. Existing fine-tuning methods using Reinforcement Learning (RL) with exact-match reward signals are often brittle, encourage memorization of training categories, and fail to elicit differential reasoning needed for generalization to unseen classes. To address this, we propose $\textbf{DiVE-k}$, $\textbf{Di}$fferential $\textbf{V}$isual r$\textbf{E}$asoning using top-$\textbf{k}$ generations, framework that leverages model's own top-k predictions as a training signal. For each training image, DiVE-k creates a multiple-choice question from the model's top-k outputs and uses RL to train the model to select the correct answer. This approach requires the model to perform fine-grained differential reasoning among plausible options and provides a simple, verifiable reward signal that mitigates memorization and improves generalization. Experiments on five standard fine-grained datasets show that our method significantly outperforms existing approaches. In the standard base-to-novel generalization setting, DiVE-k surpasses the QWEN2.5-VL-7B and ViRFT by 10.04% and 6.16% on the Harmonic Mean metric, respectively. Further experiments show similar gains in mixed-domain and few-shot scenarios. Our code is available $\href{https://github.com/raja-kumar/DiVE-k}{here}$