Repository navigation
Releases
AVL-SubtitleOCR v0.1.0 · 首个 Windows 桌面版
下载 / Download
立即下载安装包 / Download installer · v0.1.0 · 211 MB · SHA-256 校验文件 / Checksums
适用于 Windows 10 / 11(64 位),下载后双击安装。 / For 64-bit Windows 10 / 11. Download and run the installer.
AVL-SubtitleOCR v0.1.0 首次发布,为 Windows 用户提供从视频硬字幕识别、可视化校对到 SRT 导出的本地工作台。
本版本功能
- 自动识别硬字幕:检测视频画面中的字幕区域,识别文字并合并为字幕时间轴。
- 批量处理:支持多个视频加入队列,查看进度、停止或重跑任务,并批量导出 SRT。
- 可视化校对:对照视频画面修改字幕文字和起止时间,再导出标准
.srt文件。 - 多语言识别与双语界面:提供 53 个识别语言选项;界面支持简体中文和 English,首次启动默认语言跟随安装时的语言选择,也可在设置中切换。
- 本地运行:视频读取、OCR 和字幕导出在本机完成,无需上传视频;支持 CPU 推理,也可尝试通过 DirectML 使用 GPU 加速。
- 外观设置:支持浅色 / 深色主题及多种重点色。
下载与安装
- 点击顶部下载按钮,或直接下载
AVL-SubtitleOCR-Setup-0.1.0.exe。 - 双击安装包,选择安装语言和安装目录,按提示完成安装。
- 启动应用,拖入视频、选择识别语言并开始识别;完成后进入“校对”页检查结果,再导出 SRT。
系统要求:Windows 10 / 11,64 位。 无需手动安装 Python、CUDA、cuDNN 或 OCR 环境。
支持 mp4、mov、mkv、avi、webm 视频,也可拖入指向视频的 Windows 快捷方式(.lnk)。输出目录留空时,SRT 默认保存到视频旁边。
使用提示
- 本工具识别的是压制在视频画面里的硬字幕;不提供音频转写或内挂字幕轨抽取。
- 当前安装包尚未做代码签名。如出现 Windows SmartScreen 提示,请确认文件来自本仓库 Release 后再继续安装。
- GPU 加速不可用时会回落到 CPU;处理速度取决于视频长度、字幕密度、采样设置与设备性能。
- 部分视频在预览前需要本地重封装或转码,首次打开可能需要等待。
详细操作见 使用指南。遇到问题请通过 Issues 提交版本号、Windows 版本、复现步骤及脱敏后的日志。
文件校验
安装包 SHA-256:
da24c0fc966ed1d1c4e52e7c57c24372fe75721d8e8ab057dec6c80f12c49533
也可下载随附的 SHA256SUMS.txt 核对文件完整性。
English
The first release of AVL-SubtitleOCR brings local hardcoded subtitle extraction, visual proofreading, and SRT export to Windows.
- Automatically detect subtitle regions, recognize text, and build subtitle timelines.
- Queue multiple videos, review progress, proofread text and timing against the video, and export SRT files.
- Choose from 53 recognition language options and switch between Simplified Chinese and English interfaces.
- Process videos locally with CPU inference or optional DirectML GPU acceleration.
Download the Windows installer (AVL-SubtitleOCR-Setup-0.1.0.exe, 211 MB). Requires 64-bit Windows 10 / 11; no manual Python, CUDA, cuDNN, or OCR environment setup is needed.
The installer is currently unsigned. If Windows SmartScreen appears, verify that the file was downloaded from this repository's Release before proceeding. This application extracts text embedded in video frames; it does not transcribe audio or extract embedded subtitle tracks.
See the English README for more details. The SHA-256 checksum is listed above and included in SHA256SUMS.txt.