Summary of Enhancing Object Detection Performance For Small Objects Through Synthetic Data Generation and Proportional Class-balancing Technique: a Comparative Study in Industrial Scenarios, by Jibinraj Antony and Vinit Hegiste and Ali Nazeri and Hooman Tavakoli and Snehal Walunj and Christiane Plociennik and Martin Ruskowski
Enhancing Object Detection Performance for Small Objects through Synthetic Data Generation and Proportional Class-Balancing Technique: A Comparative Study in Industrial Scenarios
by Jibinraj Antony, Vinit Hegiste, Ali Nazeri, Hooman Tavakoli, Snehal Walunj, Christiane Plociennik, Martin Ruskowski
First submitted to arxiv on: 23 Jan 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper proposes a novel approach to improve object detection (OD) performance on small objects by injecting synthetic data points into the training process. This method aims to address the challenges of collecting and annotating small object data, which is time-consuming and prone to human errors. The study demonstrates the effectiveness of a simple proportional class-balancing technique for improving anchor matching in OD models. The results are compared across state-of-the-art OD models, including YOLOv5, YOLOv7, and SSD, using combinations of real and synthetic datasets within an industrial use case. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps solve the problem of object detection on small objects. Right now, it’s hard to get enough data for these small objects because it takes a long time and people make mistakes. To fix this, the researchers created new fake data points to help train the models. They also found that balancing the classes in the training data makes the models work better. The study compares some of the best object detection models, like YOLOv5 and SSD, using real and fake datasets. |
Keywords
* Artificial intelligence * Object detection * Synthetic data
