keyframe-extraction

Let Claude (or any LLM) actually watch a video — scene-aware, deduplicated frames + transcript, from a URL or local file. Runs locally, MIT.

Self-hosted AI video analysis: summaries, chapters, keyframes, transcripts and SRT / JSON export. AI 视频总结、字幕提取与结构化分析。

🎮 AI agent skill that revives shut-down mobile games as fully offline memorial builds — plus a zero-token pipeline that rescues lost game data from screenshots & screen recordings. 停服手游离线化 · 录屏抽帧OCR

Adaptive video frame extraction for SfM, Gaussian Splatting, and photogrammetry.

Generalized Keyframe Extraction for VideoQA and Video-Guided Agentic Tasks. Accepted by ECCV 2026.

Local-first MCP server for video: download any video from almost any URL, or get a transcript plus scene-aware deduplicated keyframes. YouTube, TikTok, Instagram, X, WeChat Channels, MP4/HLS. No cloud, no API keys.

本地优先的创作者素材库：视频、图片、音频与文本统一管理和检索。支持 SQLite FTS5 关键词搜索、EmbeddingGemma 2 语义搜索、文字搜图片与视频关键帧的跨模态搜索，以及中文 ASR、句级时间戳、场景检测与 AI 画面描述。Local-first creator asset search with FTS5, semantic and cross-modal retrieval.

B站 / YouTube 视频 → 图文笔记：按章节抽帧 + 字幕/转写 + 生成单篇 HTML，可投递到 IMA 等知识库

MCP server for video intelligence — analyze any video with AI vision. Extract keyframes, transcripts, and trading strategies from YouTube, Instagram & 1000+ sites.

A curated, mathematically rigorous survey and resource catalog for unsupervised, self-supervised, and multimodal foundation-model video summarization.

Compress long video into LLM-ready artifacts - semantic representative frames, timeline-aligned subtitles, motion segments. Offline file transcription + RTSP live understanding. Python, MIT.

Локальный анализ видео в Claude Code: созвоны (транскрипт + диаризация) и демо конкурентов (умный отбор кадров, привязка к речи) → PDF-отчёт. Без облака и платных API.

Fit a video into a token budget for multimodal LLMs. Picks the frames that carry the most information instead of sampling on a timer.

智能视频镜头分析工具：场景切分、关键帧提取、导演向运镜/构图/动作分析，支持成片/线稿双模式。

An end-to-end video summarization pipeline implementing and evaluating Compact BiLSTM and Interpretable Transformer architectures for keyframe selection. Features frozen MobileNetV3 feature extraction, custom Top-K duration budget policies, and performance benchmarking on the TVSum and SumMe datasets.

A unified pipeline for video frame reconstruction combining classical CV (HSV keyframes + Farneback optical flow) with deep-learning interpolation via RIFE v4.25 for high-quality, perceptually consistent frames.

15 秒至少抽 10–12 帧，把视频拆成可复刻的逐镜合同 | 10-12 keyframes per 15s: turn any video into a shot-by-shot reproduction contract · 📮 wow.vip@foxmail.com · 🎬 https://v.douyin.com/07EjIDGVErY/

Does input curation alone improve 3D Gaussian Splatting from casual video? Controlled A/B benchmark (matched budget, paired held-out views) + the evaluated frame-extraction tool. SIGGRAPH Asia 2026 Posters.

Sanitized HCMAIC 2026 E2E statistics, artifact hashes, execution status, and keyframe coverage research brief

15 个平台随便贴链接，视频秒变逐字稿 | One link turns 15+ video platforms into transcript, summary, and timeline · 📮 wow.vip@foxmail.com · 🎬 https://v.douyin.com/07EjIDGVErY/