Introduction: Indian Food Recognition Using YOLOv8 on BrainyPI

In a world where technology continues to revolutionize various aspects of our lives, it comes as no surprise that even the art of recognizing and appreciating food has found its digital counterpart.

In this blog, we unveil an innovative Indian food recognition system powered by YOLOv8, a state-of-the-art object detection algorithm. With the ability to identify a diverse range of dishes, from the aromatic Paneer Butter Masala to the rich and flavorful Biryani, this system opens doors to a new dimension of culinary exploration.

In this project we have implemented an Indian Food Recognition Model using YOLOv8 algorithm on Brainy Pi.

Supplies

  • BrainyPI
  • ShunyaOS

Step 1: Connecting to BrainyPi

Connect to the BrainyPi Remotely by SSH

ssh -X pi@auth.iotiot.in -p 65532

Enter password to establish the connection.

Step 2: Transferring Files From Local Directory to BrainyPi

Transfer the model weights, test images and inference script from local pc to brainypi

rsync -avz -e "ssh -p 65532" /home/jignesh/brainy/IFR/* pi@auth.iotiot.in:/home/pi/IFR


Enter the password to transfer the files into Brainypi


Step 3: Inference on BrainyPi


Change working directory to IFR/demo

cd IFR/demo/


Run the inference script

  python3 main.py


Inferences are saved in the output folder in current working directory.

  1. The Detected Class from the image is printed on the terminal
  2. The Image with bounding boxes around the predicted class along with the confidence of predictions are saved in the output directory


Step 4: Re-Transferring the Output Folder to the Local PC

Use the same rsync command to retransfer the output folder back to the localpc

rsync -avz -e "ssh -p 65532" pi@auth.iotiot.in:/home/pi/IFR/demo/output/* home/jignesh/brainy/output