https://github.com/reshalfahsi/anchor-free-object-detection

Tutorial on Anchor-Free Object Detection
https://github.com/reshalfahsi/anchor-free-object-detection

anchor-free computer-vision gluoncv object-detection pascal-voc

Last synced: about 6 hours ago
JSON representation

Tutorial on Anchor-Free Object Detection

• Host: GitHub
• URL: https://github.com/reshalfahsi/anchor-free-object-detection
• Owner: reshalfahsi
• Created: 2023-07-06T06:33:50.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2024-03-03T11:08:39.000Z (9 months ago)
• Last Synced: 2024-03-04T11:44:17.055Z (9 months ago)
• Topics: anchor-free, computer-vision, gluoncv, object-detection, pascal-voc
• Language: Jupyter Notebook
• Homepage:
• Size: 15.1 MB
• Stars: 1
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Anchor-Free Object Detection

 


    

    


 


Anchor boxes have been the prevalent way to generate candidates for the ground truth bounding boxes in the object detection problem. Yet, this approach is such a hassle and downright confusing. This tutorial leverages an object detection method named [FastestDet](https://github.com/dog-qiuqiu/FastestDet) that is lightweight and anchor-free. ``PASCAL VOC 2007 and 2012`` datasets are utilized to evaluate the model's capability. Here, the train and validation sets of ``PASCAL VOC 2012`` are used for the train and validation while the test set of ``PASCAL VOC 2007`` is allotted for the testing phase in this tutorial. Eventually, the inference set (the test set of ``PASCAL VOC 2007``) is used to see the qualitative performance of the model.

## Experiment

Explore [here](https://github.com/reshalfahsi/anchor-free-object-detection/blob/master/AnchorFreeObjectDetection.ipynb) to execute training, testing, and inference.

## Result

## Quantitative Result

The table below presents the quantitative result of the model on the test set.

Test metric | PASCAL VOC 2007

------------ | -------------

Loss |  3.058

[email protected]:0.95 | 15.307%

## Loss Curve

  
 Loss curve on the train set and the validation set. 


## Qualitative Result

The qualitative results of the model on the inference set are shown below.

  
 Two motorbikes (left), a person and a horse (middle), and a car and an aeroplane (right) are detected. 


## Credit

- [⚡FastestDet⚡](https://github.com/dog-qiuqiu/FastestDet)

- [GluonCV: a Deep Learning Toolkit for Computer Vision](https://cv.gluon.ai/contents.html)

- [Object Detection with RetinaNet](https://keras.io/examples/vision/retinanet/)

- [PyTorch Lightning](https://lightning.ai/docs/pytorch/latest/)

- [PASCAL VOC](http://host.robots.ox.ac.uk/pascal/VOC/)