What Would You Click? Personalized Video Thumbnail Generation with Preference-aware Highlight Retrieval

2026-07-14

AI Summary

This work addresses the lack of personalization in existing video thumbnail generation methods, which often fail to align with individual user click preferences. To bridge this gap, we introduce the first personalized video thumbnail generation task and propose a two-stage framework. In the first stage, a preference-aware highlight retriever integrates user interests with video semantics to select diverse visual anchor frames. In the second stage, a vision-language model (VLM) guides a diffusion model to generate high-fidelity, content-faithful thumbnails under precise control. Evaluated on two public datasets, our approach achieves state-of-the-art performance. User studies further demonstrate that the generated thumbnails significantly improve click-through rates and user engagement.

Abstract

Video thumbnails are a key factor for attracting user clicks on video platforms, and are increasingly supported by automation. However, existing thumbnail generation methods typically produce generic results shared across users, overlooking the diversity of individual preferences. We therefore introduce personalized video thumbnail generation, a novel task that aims to create thumbnails tailored to user-specific preferences. It is challenging in two aspects: (i) identifying visual anchors (i.e., key frames) from each video to guide the generation, which requires a balance between personalization and informativeness that existing highlight detection methods fail to achieve; and (ii) generating personalized thumbnails that are both visually coherent and faithful to the original video. As a response, we propose a two-stage framework that tightly couples preference-aware retrieval with controllable generation. In the first stage, a personalized highlight retriever captures fine-grained user-video interactions and incorporates video semantics through summarization, enabling the selection of diverse visual anchors aligned with both user preferences and video contexts. In the second stage, a VLM-guided diffusion pipeline transforms these anchors into thumbnails by extracting and injecting semantically grounded visual cues, improving personalization while preserving visual coherence and fidelity. Experiments on two public datasets show our method delivers state-of-the-art performance compared with both retrieval-based and generative baselines. A user study further demonstrates improved click preference, highlighting its effectiveness in enhancing user engagement. The code is available at https://github.com/hezy18/PVTG.

Problem

• personalized thumbnail generation
• user preference
• video highlight retrieval
• click-through rate
• visual personalization

Innovation

• personalized thumbnail generation
• preference-aware retrieval
• visual anchor selection
• VLM-guided diffusion
• user engagement