Real-Time License Plate Identification using AWS

Used models: YOLOv3 at 416px, CRAFT text detector, and CRNN text recognizer.
The number of instances required: about 20 (with cheap K80s GPUs) to achieve a smooth real-time stream. Maybe just a fraction of them if more powerful GPUs are used (T4s or V100s).
Used platform: github.com/cortexlabs/cortex.
The video has been fed to an app that then sends each frame to the cortex cluster for inferencing purposes. Even though I used a video file, it's exactly the same as if I would have used a camera as a source instead.

[@bin_jets]

This is realy cool I love it

[@nawt12]

using toast machine to take this video :D on the other hand very usefull i think can be more attractive with HD video. good job.

[@albost1]

Is there a repository for this?, where should I start to make a project like this one in python?

[@bin_jets]

It's you who build it?

[@alexpetro4252]

Not me but will learn it next time

[@bin_jets]

@@alexpetro4252 Please learn it to us also will you 🙏

[@enver.suarez]

el OCR esta del asco