AI Networking Cookbook: Practical recipes for AI-assisted network automation and development
TL;DR
A Motion-Aware Frame Sampling (MAFS) strategy for constructing motion-aware captions enables more effective utilization of video frames and improves captioning quality for large-motion videos.
ClearText-Video (CTVid), a large-scale, scene-text-aware benchmark for studying text-centric video understanding under controlled quality variation, is introduced, providing a rigorous foundation for restoration-aware, quality-robust text-centric video systems.
This work proposes a large-scale multimodal large language model-based supervision generation framework that improves supervision diversity, fidelity, and semantic coverage, and introduces a granularity-aware text representation with separate CLS tokens for summary and detailed views.
No-reference video quality assessment (NR VQA) has recently seen promising progress with deep learning. However, video data is inherently large, and processing them with deep models incurs high computational cost. This challenge is particularly acute in VQA, where preserving original-resolution cues and dense temporal...
ViTAL-X, a lightweight model that equips frozen image-text backbones with temporal awareness while preserving their foundational spatial knowledge, achieves state-of-the-art performance and demonstrates that targeted, high-quality temporal alignment provides a highly efficient alternative to pure scaling.
This work introduces StrAD, a benchmark for long-form AD generation on full-length videos spanning diverse genres such as movies, documentaries, short films, performances, and video games, and reformulates AD generation as streaming dense video captioning, generating ADs on the fly without ground-truth timestamps.
This work presents LAION-BVD, a large-scale open video dataset for multimodal learning, which contains 1.3B platform-specific video URLs collected from CommonCrawl, and explores video frames as an alternative source of image-text data by extracting scene-changing frames.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.