Stroma Vision Task
We are building a computer vision pipeline for a nuts-and-bolts manufacturer. For Dataset operations, preprocessing operations were carried out with the notebook locally. For Training and Testing, object detection with Yolov8 and tracking with ByteTrack & Supervision are performed in Colab.
Getting Started
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.
Prerequisites
What things you need to install the software and how to install them
Installing
Create Python Virtual Environments
Since notebook will be used in local for dataset operations, virtual environment is created with the following commands.
conda create --name stroma-vision python=3.10.6
pip install -r requirements.txtRun
git clone https://github.com/Tolgaisikp/stroma-vision-task.gitYou can convert the data given coco annotation to yolo annotation by running the first 3 notebooks above. to mention the above notebooks respectively
- 1-extract-image.ipynb : This file contains functions to extract pictures from train test and val videos in the challenge folder.
- 2-coco-to-yolov8.ipynb : This file converts the dataset which is coco annotation to yolo annotation with pylabel importer and creates the dataset.yaml file.
- 3-delete-empty-img.ipynb : This file deletes the empty labels by looking at the sizes of the labels created in 2-coco-to-yolov8 and accordingly deletes the images of the deleted labels.
- 4-custom-yolov8-train_and_test.ipynb : This file contains the Google Drive link of the dataset for Training and Testing, the Github link of the code and the markdown containing the Colab link.
Built With
- Colab - GPU powered Jupyer Notebook
- YoloV8 - YoloV8 Github link
- Dataset - Latest version of the Dataset
Screenshots
Authors
Tolga Işık - Tolgaisikp
License
This project is licensed under the MIT License - see the LICENSE.md file for details