ShotBench: Expert-Level Cinematic Understanding in Vision-Language Models
This addresses the problem of limited fine-grained visual comprehension for AI-assisted video generation, though it is incremental as it builds on existing VLM methods with new data and optimization.
The paper tackles the lack of robust evaluation for cinematic understanding in Vision-Language Models (VLMs) by introducing ShotBench, a benchmark with over 3.5k expert-annotated QA pairs, revealing that top models achieve less than 60% accuracy. They also develop ShotVL using a 70k-pair dataset, which significantly outperforms existing models and establishes new state-of-the-art performance.
Cinematography, the fundamental visual language of film, is essential for conveying narrative, emotion, and aesthetic quality. While recent Vision-Language Models (VLMs) demonstrate strong general visual understanding, their proficiency in comprehending the nuanced cinematic grammar embedded within individual shots remains largely unexplored and lacks robust evaluation. This critical gap limits both fine-grained visual comprehension and the precision of AI-assisted video generation. To address this, we introduce ShotBench, a comprehensive benchmark specifically designed for cinematic language understanding. It features over 3.5k expert-annotated QA pairs from images and video clips, meticulously curated from over 200 acclaimed (predominantly Oscar-nominated) films and spanning eight key cinematography dimensions. Our evaluation of 24 leading VLMs on ShotBench reveals their substantial limitations: even the top-performing model achieves less than 60% average accuracy, particularly struggling with fine-grained visual cues and complex spatial reasoning. To catalyze advancement in this domain, we construct ShotQA, a large-scale multimodal dataset comprising approximately 70k cinematic QA pairs. Leveraging ShotQA, we develop ShotVL through supervised fine-tuning and Group Relative Policy Optimization. ShotVL significantly outperforms all existing open-source and proprietary models on ShotBench, establishing new state-of-the-art performance. We open-source our models, data, and code to foster rapid progress in this crucial area of AI-driven cinematic understanding and generation.