Object Detector

Description

This is a simple webpage to allow any user a drag and drop interface to get predictions from yolov5.

The image addition can be done in a drag and drop, or file selection manner.

It will show the Image, the bounding boxes and the list of objects detected.

Packages used

Front end

• React as a framework for building the webpage
• Konva to draw the image and the bounding boxes
• Mui for components

Back end

• Tornado to build the Api
• Pytorch to load the model and get predictions

Container

• Docker

Running the Service

Download the code and run

bash build.sh

then you can access the webpage at http://localhost:80