Comments (2)
Hi @Kraufel,
Thank you for reporting this issue. I have already fixed the problem you mentioned.
Regarding your second question, to make it clearer, I also changed the names of some the formats of the "detection in the interface". So, please update the files with these changes.
Find below the explanation why if you use your coordinates, it wont work.
Considering your groundtruth:
0 0.1466346153846154 0.4230769230769231 0.09134615384615384 0.08173076923076923
These coordinates are in YOLO's format: <class_id> <x_center> <y_center> <width> <height> (RELATIVE)
And your detection: 0 0.74 0.271635 0.153846 0.687500 0.427885
is in the format: <class_id> <confidence> <x_center> <y_center> <width> <height> (RELATIVE)
If you take a look at the paper (page 6), you will note that YOLO's coordinates are represented in the format:
As the coordinates are normalized by the image's resolution, you have to inform the directory with the images, so the program can retrieve the width and height of the image to calcualte the coordinates of the bounding boxes.
If the groundtruths and detections above belong to the image 2007_000549.jpg, whose width and height are 375 and 500 respectively, their absolute coordinates in the format x, y, x2, y2 will be:
Groundtruth: (38, 191, 72, 232)
Detection: (-27, -30, 231, 184)
See that the x and y coordinates of the detections are negative, which means this detection cannot belong to this image.
I don't know where you got that detected bounding box, but it cannot belong to the image 2007_000549 in the toyexample.
I hope it helps. Let me know if you were able to run the program with your files.
Best regards,
Rafael
from review_object_detection_metrics.
As no activity has been performed in this issue, I decided to close it.
If anyone decides to add any comment or have doubts on this topic, feel free to reopen the issue.
from review_object_detection_metrics.
Related Issues (20)
- Equation (9) and (11) in publication HOT 3
- how to use commond line ? HOT 1
- Do you have the detection results of different detection models in coco json format. HOT 1
- could't open detected the annotation file made by yolov4 darknet HOT 2
- How to calculate the general Precision x recall curve of the detection model? HOT 1
- Coordinates format has repeated option HOT 4
- Can you share commands on how to run the coco and PascalVOC evaluators via terminal to generate the outputs? HOT 3
- No results HOT 4
- Feature request: Docker containerization HOT 3
- move(self, int, int): argument 1 has unexpected type 'float' HOT 5
- ModuleNotFoundError: No module named 'src' HOT 4
- Bug in the converter.py while looking for images folder ? HOT 1
- AttributeError: 'list' object has no attribute 'items' HOT 1
- Please provide command line to evaluate on terminal HOT 1
- Issue during start: TypeError: arguments did not match any overloaded call HOT 3
- np.bool is deprecated HOT 5
- numpy: set_window_title moved to manager HOT 3
- Building the conda environment causes an endless loop HOT 2
- Result is 0 HOT 3
- Getting precision, recall, F1 values 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 review_object_detection_metrics.