CropVLM: Learning to Zoom for Fine-Grained Vision-Language Perception

arXiv:2511.19820v2 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) often struggle with tasks that require fine-grained image understanding, such as scene-text recognition or document analysis, due to perception limitations and visual fragmentation. To address these challenges, we introduce CropVLM as an external low-cost method for boosting performance, enabling VLMs to dynamically ''zoom in'' on relevant image regions, enhancing their ability to capture fine details. CropVLM is trained using reinforcement learning, without using human-labeled bounding boxes as a supervision signal, and without expensive synthetic evaluations. The model is trained once and can be paired with both open-source and proprietary VLMs to improve their performance. Our approach delivers significant improvements on tasks that require high-resolution image understanding, notably for benchmarks that are out-of-domain for the target VLM, without modifying or fine-tuning the VLM, thus avoiding catastrophic forgetting.

Related

• Q-Zoom: Query-Aware Adaptive Perception for Efficient Multimodal Large Language Models
• VisionFoundry: Teaching VLMs Visual Perception with Synthetic Images
• Lang2Act: Fine-Grained Visual Reasoning through Self-Emergent Linguistic Toolchains
• Boosting Visual Instruction Tuning with Self-Supervised Guidance
• Growing a Multi-head Twig via Distillation and Reinforcement Learning to Accelerate Large Vision-Language Models

Source: arXiv cs.AI | 2026-04-15