Having the server decide the image format based on the accept header is simpler. Services like https://imagekit.io/ (no affiliation) can do that for you. - Source: Hacker News / about 2 months ago

Hosting wise, I would reccomend pythonanywhere.com, combined with either https://imagekit.io or https://cloudinary.com. Source: about 1 year ago

Use any third-party service to store images like Cloudinary , imagekit etc... And store image URLs in your database. If you have fewer images you keep image URLs directly in your APIs. Source: over 1 year ago

Imagekit.io – Image CDN with automatic optimization, real-time transformation, and storage that you can integrate with existing setup in minutes. Free plan includes up to 20GB bandwidth per month. - Source: dev.to / over 1 year ago

We'll be using Multer to handle file uploads on our server and imagekit will do all our media heavy lifting. I chose these tools because I just found them easier to use and the latter has a very elaborate documentation (and a free tier too 😋). - Source: dev.to / over 1 year ago

OpenCV and "AI" can work well together; see YOLO: https://pjreddie.com/darknet/yolo/. - Source: Hacker News / 6 months ago

Then there is the creator of YOLO. His resume is epic. It's completely My Little Pony themed. Source: over 1 year ago

For the API, I've used python and django. For image processing and detecting persons in image, I used yolov3. If any person exceeded limit that user gave, the API sends notification to user via telegram. Source: almost 2 years ago

The paper says the source code is available: Https://pjreddie.com/darknet/yolo/. Source: about 2 years ago

If you are trying to identify objects in an image (like 'cat', 'person', etc), then you might look at YOLO. There are lots of github projects using it to explore, too. Source: about 2 years ago