Training a Computer Vision Model for Commercial Bakeries with Primarily Synthetic Images

TL;DR

This paper develops a computer vision system for bakery product detection using a large dataset and synthetic images, achieving high accuracy with state-of-the-art models.

Contribution

It introduces an expanded dataset and synthetic image generation to improve bakery product detection with YOLO models.

Findings

Achieved 90.3% AP@0.5 accuracy on test set.

Synthetic images effectively enhanced model robustness.

Extended previous work with more diverse data.

Abstract

In the food industry, reprocessing returned product is a vital step to increase resource efficiency. [SBB23] presented an AI application that automates the tracking of returned bread buns. We extend their work by creating an expanded dataset comprising 2432 images and a wider range of baked goods. To increase model robustness, we use generative models pix2pix and CycleGAN to create synthetic images. We train state-of-the-art object detection model YOLOv9 and YOLOv8 on our detection task. Our overall best-performing model achieved an average precision AP@0.5 of 90.3% on our test set.