Add to .env file "max_detect_size" option

To improve performance on huge images, we resize it before face detection. The downside of this is that small faces could be not recognized.

Some users complain that we do not catch too small faces in their 1080p camera.

We need to make resizing a configurable variable, so users choose if we need to resize at all, or what resize to apply

Here is where we init the variable in Python: https://github.com/exadel-inc/CompreFace/blob/c44c92eaea31798f8e68c91acc36144d3b8ab411/embedding-calculator/src/constants.py#L24

So basically all we need is to add the “IMG_LENGTH_LIMIT” environment variable to docker-compose and then add it to .env file

[pospielov]

Hi, thanks for the interest in our repository, I assigned the issue to you

[pospielov]

Merged, thanks for the contribution!