3

From a research study involving participants with a camera, I have chunks of video recording. I have to split them as part of the preprocessing into short sequences of around 5 min long. Each file will be named after its time-stamp in the sequence. I can store them as binary blobs in a database.

In this study, I also have other sensors. It's clear that I should Use a time series DBMS to store these data (IoT sensing data). This will facilitate queries based on time. Do you see a problem with me storing the file paths and indexing them into a TimeseriesDB (like InfluxDB) based on their file blob's sequential time-stamps to facilitate time-driven queries?

The Q & A Should binary files be stored in the database? partially answers my question - from looking at that answer is sounds like storing video as a blob is a good solution. But my question is more about indexing video by the file's time-stamps.

The video is recorded in large (time) chunks at a time. as part of the cleaning/pre-processing, we split them into 5-minute long segments. Each segment is saved to a file that retains/maintains a date-time-stamp (a 50 min long video is split into 10). Putting it this way, it seems natural to store these file paths per their date-time-stamp. I just want to run my logic by others.

Improve this question
0

2 Answers 2

Reset to default
0

From a research study involving participants with a camera, I have chunks of video recording. I have to split them as part of the preprocessing into short sequences of around 5 min long. Each file will be named after its time-stamp in the sequence. I can store them as binary blobs in a database.

It doesn't make much sense to split apart a video, unless for security. It's a container with an audio and video stream. Normally they're stored in a container like AVI, MP4 or MKV (the open standard). In MKV-parlance, you have "Chapters". This is effectively a timeseries database.

Also, you may want to be careful if you're extracting and cutting them. Keep in mind most video streams are compressed, and de-compressing and re-compressing can introduce artifacts and such.

Here are the fields store in the Chapter's data

See Also

Improve this answer
1
  • 1
    For privacy, we want to offer the participant a chance to review the video and choose the parts they are not willing to share with researchers. By splitting it, we can easily delete (mark for deletion, or label as not-for-viewing) those particular 5 min segments.
    – sAguinaga
    Jan 7, 2019 at 1:32
0

I tried to use InfluxDB for keeping timestamps and meta information about photos from a CV camera and a file system for the data. It works fine if you don't need to remove old data, then you run out of disk space.

If you need something like a quota for your data, try ReductStore.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.