IoU threshold influence on detecting results when using detect.py

fyang5 commented 8 months ago

Search before asking

[X] I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

I trained the yolov5n model with a customized dataset, and then test the testing data with detect.py using the command: python detect.py --source /test/videos --weights runs/train/yolov5n/exp/weights/best.pt --save-txt --device 0 --conf-thres 0.2 --iou-thres 0.2 --save-conf .

However, when I fix the confidence score =0.2, then using make IoU threshold varies from 0.1 to 0.9, I found that the results does not change at all. In contrast, if I fix the IoU threshold=0.1, then let confidence score vary from 0.1 to 0.9, then results decreases as expected. See the figure below. I cannot understand why the IoU threshold does not change the result. I would expect the result decreases as confidence score. Is there something wrong?

Additional

No response

github-actions[bot] commented 8 months ago

👋 Hello @fyang5, thank you for your interest in YOLOv5 🚀! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a 🐛 Bug Report, please provide a minimum reproducible example to help us debug it.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset image examples and training logs, and verify you are following our Tips for Best Training Results.

Requirements

Python>=3.8.0 with all requirements.txt installed including PyTorch>=1.8. To get started:

git clone https://github.com/ultralytics/yolov5  # clone
cd yolov5
pip install -r requirements.txt  # install

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

Notebooks with free GPU:
Google Cloud Deep Learning VM. See GCP Quickstart Guide
Amazon Deep Learning AMI. See AWS Quickstart Guide
Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training, validation, inference, export and benchmarks on macOS, Windows, and Ubuntu every 24 hours and on every commit.

Introducing YOLOv8 🚀

We're excited to announce the launch of our latest state-of-the-art (SOTA) object detection model for 2023 - YOLOv8 🚀!

Designed to be fast, accurate, and easy to use, YOLOv8 is an ideal choice for a wide range of object detection, image segmentation and image classification tasks. With YOLOv8, you'll be able to quickly and accurately detect objects in real-time, streamline your workflows, and achieve new levels of accuracy in your projects.

Check out our YOLOv8 Docs for details and get started with:

pip install ultralytics

fyang5 commented 8 months ago

I also tested with Yolov8, and it generated the same results.

glenn-jocher commented 8 months ago

@fyang5 hello! The IoU (Intersection over Union) threshold is used during Non-Maximum Suppression (NMS) to determine whether two detected bounding boxes overlap too much and are likely to be the same object. If the IoU threshold does not influence your results, it could be due to a few reasons:

Sparse Detections: If your detections are sparse and there's little overlap between bounding boxes, changing the IoU threshold won't have much effect.
High Confidence Detections: If most detections have high confidence, they might not be affected by the IoU threshold, especially if there's not much overlap.
Dataset Characteristics: The nature of your dataset might be such that objects are not close enough to create overlaps that would be affected by the IoU threshold.

It's also worth noting that the IoU threshold is more influential in crowded scenes where multiple objects are close together. If your scenes are not crowded, the IoU threshold will have less impact.

For a more detailed analysis, you might want to visually inspect the detections at different IoU thresholds or check the distribution of IoU scores for your detections.

If you continue to experience unexpected behavior, please ensure you're using the latest version of the code and consider opening an issue with detailed steps to reproduce the problem. For further guidance, you can refer to our documentation at https://docs.ultralytics.com/yolov5/.

Keep in mind that YOLOv8 is a separate entity and might have different behaviors or settings that could influence the results in ways not covered here.

Happy detecting! 😊🚀

github-actions[bot] commented 7 months ago

👋 Hello there! We wanted to give you a friendly reminder that this issue has not had any recent activity and may be closed soon, but don't worry - you can always reopen it if needed. If you still have any questions or concerns, please feel free to let us know how we can help.

For additional resources and information, please see the links below:

Docs: https://docs.ultralytics.com
HUB: https://hub.ultralytics.com
Community: https://community.ultralytics.com

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLO 🚀 and Vision AI ⭐

ultralytics / yolov5