【ML Paper】Explanation of all of YOLO series Part 8

This is an summary of paper that explains yolov1 to v8 in one.
Let's see the history of the yolo with this paper.
This article is part 8, part 7 is here.

Original Paper: https://arxiv.org/pdf/2304.00501

4.4 YOLOv1 Strengths and Limitations

High-Speed Performance

YOLO's simple architecture, combined with its novel full-image one-shot regression, enables it to process images significantly faster than existing object detectors. This efficiency allows YOLO to achieve real-time performance, making it highly suitable for applications requiring rapid object detection.

Increased Localization Error

Despite its speed advantages, YOLO experiences larger localization errors compared to state-of-the-art methods such as Fast R-CNN. This limitation is primarily due to three factors:

Firstly, YOLO is restricted to detecting at most two objects of the same class within a single grid cell, which hampers its ability to accurately predict objects that are in close proximity. Secondly, the model struggles to predict objects with aspect ratios that were not represented in the training data, limiting its adaptability to diverse object shapes. Lastly, the use of down-sampling layers causes YOLO to learn from coarse object features, reducing the precision of object localization.