Semmeldetector: Application of Machine Learning in Commercial Bakeries

Thomas H. Schmitt,Maximilian Bundscherer,Tobias Bocklet

2024-06-06

Abstract:The Semmeldetector, is a machine learning application that utilizes object detection models to detect, classify and count baked goods in images. Our application allows commercial bakers to track unsold baked goods, which allows them to optimize production and increase resource efficiency. We compiled a dataset comprising 1151 images that distinguishes between 18 different types of baked goods to train our detection models. To facilitate model training, we used a Copy-Paste augmentation pipeline to expand our dataset. We trained the state-of-the-art object detection model YOLOv8 on our detection task. We tested the impact of different training data, model scale, and online image augmentation pipelines on model performance. Our overall best performing model, achieved an AP@0.5 of 89.1% on our test set. Based on our results, we conclude that machine learning can be a valuable tool even for unforeseen industries like bakeries, even with very limited datasets.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

This paper introduces a machine learning application called "Semmeldetector" which utilizes object detection models (such as YOLOv8) to detect, classify, and count products in baked food images. Due to the diverse range of baked food in Germany, there are not enough datasets to differentiate between different types of baked goods. Therefore, the researchers created a dataset consisting of 1151 images that distinguish 18 different types of baked goods. To augment the training data, they employed a data augmentation method called Copy-Paste.

The goal of Semmeldetector is to assist commercial bakeries in tracking unsold baked goods to optimize production and improve resource efficiency. Through experiments with different training data, model sizes, and online image augmentation methods, the researchers found that even with a small dataset, machine learning can provide value in unexpected fields like the baking industry. Their best model achieved an average precision (AP0.5) of 89.1% on the test set.

Furthermore, the paper emphasizes the importance of introducing random background images when training object detection models in real-world scenarios to reduce false positives and improve model performance. Finally, they propose future work could involve comparing more types of object detection models, exploring generative models for synthesizing baked food images, and expanding the training data to cover a wider variety of baked goods.