Comments (1)
@Buckler89 hey there 👋, thanks a lot for such a detailed report! You've pointed out a very valid concern regarding the unintended alteration of bounding box formats within the bbox_areas
property.
Indeed, the current design where Bboxes.areas()
changes the bounding box format to XYXY without reverting back could lead to unexpected behaviours, especially in pipelines where the original format is essential for subsequent processes.
To address this issue, a potential solution might involve modifying the Bboxes.areas()
to perform format conversion and area computation without altering the original format state of the bounding boxes. Here's a quick idea of how that could be structured:
def areas(self, revert_format=True):
"""Return box areas, optionally reverting to original format."""
original_format = self.format
self.convert("xyxy")
areas = (self.bboxes[:, 2] - self.bboxes[:, 0]) * (self.bboxes[:, 3] - self.bboxes[:, 1])
if revert_format:
self.convert(original_format)
return areas
This modification suggests adding a parameter to control whether the bounding boxes should revert to their original format after computing the areas. By default, it reverts to maintaining the original behavior unless specified otherwise.
I'll forward this to our team to verify and implement a fix efficiently. Thanks again for catching this and for your detailed explanation! 😄 If there's anything more you'd like to discuss or suggest, feel free to share!
from ultralytics.
Related Issues (20)
- Instance segmentation custom model training HOT 2
- Fine-tuning YOLOv8n Model on Two New Classes HOT 2
- Exception with YOLOWorld ONNX export in dynamic mode HOT 3
- How cna I validate my model with no-square 【640*384(w*h)】image size? HOT 3
- RuntimeError:An attempt has been made to start a new process before the current process has finished its bootstrapping phase. HOT 5
- Bug report: FileNotFoundError due to special characters in pathnames at /ultralytics/data/base.py HOT 3
- Yolov9 pytorch model is empty, as in not the folder, the loaded pytorch model. HOT 1
- Model exported to CoreML showing 1.0 confidence values for all detections HOT 4
- YOLOv8-OBB learns 0° rotation HOT 12
- train with arg use_ray=True, raise exception OSError: [WinError 87], at _winapi.CreateProcess HOT 3
- New network structure with old weights HOT 3
- For increasing or decreasing video streams,Is there a better way to detect the need to interrupt and then load new data and restart? HOT 1
- Is YOLOv8 Using Val Set to train? HOT 2
- Can YOLOv8 Simultaneously Support Rotated Object Detection (OBB) and Keypoint Detection? HOT 3
- yolov8 when train,raise error OSError: [WinError 87] parameters error HOT 1
- Will you support YOLOv10 in the future? HOT 2
- YOLOv8 OBB HOT 1
- YOLOv8 with KAN HOT 2
- A question about validation set drawing image results HOT 2
- Determine the Class of a Specific Pixel-Coordinate from YOLOv8 Segmentation Results HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from ultralytics.