🤖 AI Summary
This study addresses the challenge of detecting small, densely packed, and overlapping objects in industrial recycling scenarios by introducing the first dedicated small-object detection dataset, comprising over 10,000 images and 120,000 annotated instances. The authors systematically evaluate the performance of YOLO-family models across three tasks: small object detection, length measurement, and anomaly detection. To enhance robustness to scale variations, they propose an anomaly detection method leveraging high-resolution inputs, scale-robust data augmentation, and synthetic image generation. Experimental results demonstrate that the selected optimal YOLO variant achieves superior accuracy, efficiency, and stability, offering a reliable solution for automating industrial recycling processes and establishing a benchmark for future research in this domain.
📝 Abstract
In this paper, we address the problem of detecting small, dense, and overlapping objects, a major challenge in computer vision. Our focus is on reviewing proposed methods based on deep learning supervised approaches. We provide a detailed comparison of these systems on a new dataset of more than 10k images and 120k instances, highlighting their performance, accuracy, and computational efficiency in the industrial recycling process use case. Through this comparative analysis, we identify the most reliable systems currently available and the specific challenges they are designed to tackle. Furthermore, we explore the benefits of data augmentation and synthetic images. Based on our analysis, we also propose potential future directions and innovative solutions that could enhance the effectiveness of small, dense and overlapped object detection systems. The scope of our investigations encompasses object detection, length measurement, and anomaly detection within the context of the recycling process. The anomaly detection strategy is robust against variations in image resolution and zoom levels, ensuring reliable performance in industrial applications. The repository of the proposed dataset, methods and evaluation codes can be found at: https://github.com/o-messai/SDOOD