存储视频和索引内容

发布于 2024-09-02 09:50:10 字数 513 浏览 2 评论 0原文

如果有人提出建议,请重新审视这一点:

我被要求创建或查找满足以下要求的系统...

1) 允许上传视频文件 - 当前解决方案:PHP 上传到 Web 根目录上方的目录,然后我将在允许访问流之前验证用户。

2) 使用有关视频中参与者的元信息标记文件(这些是调查)以供以后搜索。 - 当前解决方案:关键字文本区域用“;”分隔项目然后将项目解析到数据库中的“关键字”表中以供稍后搜索。

3) 转录文本以供稍后进行全文搜索,因此,如果参与者声明“我喜欢游泳、骑自行车、跑步”,稍后搜索“跑步”将找到此结果(铁人三项可能已输入到元字段中) ) - 目前的解决方案:使用服务进行转录。然后在全文索引字段中上传文本。

这些信息将向订阅的客户提供,因此在上面的示例中,从事游泳、骑自行车、跑步的公司可能能够获得此结果,但冰淇淋供应商可能无法获得。 - 当前的解决方案:在注册和签入阶段为订阅者和视频分配类别。确保它们匹配。

似乎会有很多手动设置,所以如果有人对自动化或控制有更好的想法,请告诉我。

感谢您的建议。

Revisiting this in case someone has a suggestion:

I've been asked to either create or find a system that meet the following requirements...

1) Allow upload of video files - Current solution: PHP to upload to a directory above the web root, and then I'll verify users before allowing access to stream.

2) Tag the files with meta info about the participant in the video (these are surveys) for later searching. - Current solution: keyword text area separating items by ";" then parsing items into a "keywords" table in DB for searching later.

3) Transcribe the text for a full text search later, so if the participant states, "I like to swim, bike, run" later a search for "run" would find this result (triathlon would have probably been entered in the meta fields) - Current solution: use service to do transcriptions. Then upload text in a full text indexed field.

The info will be made available for clients with subscriptions, so in the example above, companies that deal with swimming, biking, running may be able to get this result, but ice cream vendors may not. - Current solution: assign categories to subscribers as well as videos during the sign up and check in phases. Make sure they match.

Seems like there will be a lot of manual setup, so if anyone has any better ideas for automating or controlling, please let me know.

Thanks for suggestions.

如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

扫码二维码加入Web技术交流群

发布评论

需要 登录 才能够评论, 你可以免费 注册 一个本站的账号。

评论(1

迷你仙 2024-09-09 09:50:10

简单的!构建一个单独的(简单的)标记界面,向用户呈现视频和用于更新标记的输入字段。通过亚马逊的 Mechanical Turk 让人们对单词进行标记。便宜、简单、快捷。据我所知,目前还没有服务器端解决方案来完成本质上数据库捕获的隐藏式字幕。甚至电视台也有人边听边打字。

顺便说一句,用 ids 链接的表中的各个记录可能会更好地为您的逗号分隔解决方案提供服务。不要害怕表连接,它们比逗号分隔的搜索更快更容易。

Easy! Build a separate (simple) interface for tagging that presents the user with a video and an input field to update tagging. Go through Amazon's Mechanical Turk to get people to do tagging of words. Cheap, easy, quick. Best I know, there's currently no server-side solution for doing what's essentially database-captured closed captioning. Even the television stations have people listening and typing away.

By the way, your comma separated solution might be better served by individual records in a table linked by ids. Don't fear table joins, they can be faster and easier than comma delimited searches.

~没有更多了~
我们使用 Cookies 和其他技术来定制您的体验包括您的登录状态等。通过阅读我们的 隐私政策 了解更多相关信息。 单击 接受 或继续使用网站,即表示您同意使用 Cookies 和您的相关数据。
原文