https://github.com/tanmaypandey7/wheat-detection

Detecting wheat heads using YOLOv5
https://github.com/tanmaypandey7/wheat-detection

flask kaggle-competition wheat-detection yolov5

Last synced: 12 days ago
JSON representation

Detecting wheat heads using YOLOv5

• Host: GitHub
• URL: https://github.com/tanmaypandey7/wheat-detection
• Owner: tanmaypandey7
• License: gpl-3.0
• Created: 2020-08-08T19:16:16.000Z (almost 4 years ago)
• Default Branch: master
• Last Pushed: 2021-10-18T07:52:16.000Z (over 2 years ago)
• Last Synced: 2024-02-29T10:34:06.026Z (4 months ago)
• Topics: flask, kaggle-competition, wheat-detection, yolov5
• Language: Python
• Homepage: https://colab.research.google.com/drive/1bZe1CDa4g7wnOUZFO9ZUmw15T2I6Po0w?usp=sharing
• Size: 2.69 MB
• Stars: 11
• Watchers: 2
• Forks: 5
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Lists

• awesome-yolo-object-detection - tanmaypandey7/wheat-detection - detection?style=social"/> : Detecting wheat heads using YOLOv5. (Applications)

README

        
# Wheat Detection

Detecting wheat heads using YOLOv5

- [Web App demo](#Web-app-demo)

- [Brief overview of the competition images](#Brief-overview-of-the-competition-images)

- [Modifications](#Modifications)

- [Training](#Training)

- [Inference and Deployment](#Inference-and-Deployment)

## Web App Demo

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1bZe1CDa4g7wnOUZFO9ZUmw15T2I6Po0w?usp=sharing)

[![https://imgur.com/a/Ap2kaeX](http://img.youtube.com/vi/JrL8nsV53tc/0.jpg)](http://www.youtube.com/watch?v=JrL8nsV53tc "Wheat App")

## Brief overview of the competition images

Wheat heads were from various sources:  

  

A few labeled images are as shown: (Blue bounding boxes)  

   

## Pre-trained models

Models can be downloaded from here. (Use last_yolov5x_4M50fold0.pt for best results) 

## Modifications

The YOLOv5 notebook internally does some augmentations while preparing a Dataset. 

Originally, this Dataset consists of only Mosaic images.

**Mosaic** - https://arxiv.org/pdf/2004.12432.pdf  

      4 images are cropped and stitched together  

    

    

Here, I modified the repo to add Mixup.

**Mixup** - https://arxiv.org/pdf/1710.09412.pdf  

      2 images are mixed together

    

 

I modified the code(especifically utils.datasets) so it had a 50-50 chance of creating a mixup or a mosaic image. This was very helpful for us as it boosted our public score from 0.77->0.7769. 

These developments were made before we found out that YOLOv5 was non-compliant and had to switch to EfficientDet for our final 2 submissions.

Kaggle later updated the leaderboard with the final 2 submissions and we ended up at 113th Private(Top 6%).

## Training

We trained the model for 50 epochs on Colab Pro. 

## Inference and Deployment

Our best model is currently being used for inference in this web-app. I uses HTML and CSS as front-end and Flask as the backend.

This web-app is served on Google Colab but can be easily deployed on AWS or GCP as well.