People focusing camera is very important on some senario like face-to-face e-learning as meta-data. Face API support detect several eye point and hopefully Video Indexer calcrate summary it like "camera focusing time".
https://docs.microsoft.com/en-us/azure/cognitive-services/face/face-api-how-to-topics/howtodetectfacesinimage

Too low accuarcy that Video Indexer can't detect its language with auto-detect mode like this:
https://www.youtube.com/watch?v=kRQnArIIffw
It's Japanese but Video Indexer defines this as en-us.

Is video indexer available for OnPrem Installations ? or is it only cloud ?

VI returns timecodes with hundredths of seconds (i.e. 00:00:07.2400000, so 7 and 24 hundredths of seconds). If we know the video frame-rate (i.e. 24, 23.98), could we derive the specific frame by multiplying the decimal times the frame rate? Or would that not be reliable to convert to HH:MM:SS:FF?

lots of the AI can also be used for photos. This will be very usefull in one solution (all you Video and photo media in one engine!)

Please add support for Ukrainian audio transcription.

When we embed the insights widget and a user edits it, our app needs to know when an edit has been made. Perhaps this can be surfaced through a webhook that can be specified as part of the integration code.

The Video keep loading. Forever and ever

I am using Video Indexer. I am having some suggestions below :

• The description of Search API is to be more clear. Currently we don’t understand what it searches and what would be output.
• If the Face API group ID is provided as input, auto update of face names should happen from Face Group ID.
• The sections (people, brand, keywords…) inside Embedded Insight widget could be collapsible Control.

If we were able to train the Indexer with more functionality regarding Faces and Brands we will probably move to Microsoft from Google AI

Analyze each frame or one in x frames fully for the 7 emotions your old face emotion recognition gave.
Return a rolling average of the emotion data as well as JSON.

The idea is that even if individual frames don;t give you great data a group may act as boosting and give better results.

Is CORS enabled for Video Indexer?

so far, only one language displays on screen. since we have Microsoft Translator, could we show two languages of transcripts in the meanwhile? like video is in English, we could see transcripts both in English and Chinese?

Would it be possible to extract text / numeric information and save it to a file? example video of stock market, with continuous numeric and text data, and export content to downloadable file

When the autogenerated transcript is not correct you are able to edit the words, but there is no way to adjust the timing to make those words be in sync with the video. Please enable the ability to adjust the timing.

I uploaded a video and its stuck at 65%. Any one can help here?

The ability to train the API to detect scenes and objects(faces, logos) according to my needs

Add the ability to get the XY co-ordinates of tracked people

I've done a really good test from a french tv documentary. It will be very usefull to select différent language for subtitle, instead of generate only subtitle in the native language of the media.
Thanx