search

boostcampaitech7 / level2-objectdetection-cv-10

level2-objectdetection-cv-10 created by GitHub Classroom
0 stars 0 forks source link

readme

Object Detection for classification of recycled items

πŸ“˜Overview

2024.10.02 ~ 2024.10.24

This project focuses on detecting objects in recycling trash images as part of a private competition organized by Naver Connect Foundation and Upstage.

πŸ“˜Contributors

κΉ€κΈ°μˆ˜ 문채원 μ•ˆμ£Όν˜• μ€μ˜μ°¬ μ΄μž¬ν›ˆ μž₯μ§€μš°






πŸ“˜Wrap up Report

Here's our link

πŸ“˜Metrics

https://user-images.githubusercontent.com/64190071/164357745-4d03deb3-6104-4706-a890-3d002a904067.png

https://user-images.githubusercontent.com/64190071/164357754-718a8628-872e-4f1e-9d12-4e212b2444ab.png

https://user-images.githubusercontent.com/64190071/164357763-9d7c667a-2c5a-4b92-b4ae-6c32be0b7d34.png

πŸ“°Tools

πŸ“°Folder Structure


β”œβ”€β”€ mmdetection2
β”‚   β”œβ”€β”€ configs
β”‚   β”œβ”€β”€projects/configs/custom
β”‚   β”œβ”€β”€ inference_wbf.py
β”‚   β”œβ”€β”€ train.py
β”‚   └── etc
β”œβ”€β”€ mmdetection3
β”‚   └── config files
β”œβ”€β”€ Ultralytics
β”‚   β”œβ”€β”€ RT_DETR.ipynb
β”‚   β”œβ”€β”€ RT_DETR_WBF_infernece.ipynb
β”‚   β”œβ”€β”€ Yolov11.ipynb
β”‚   └── Yolov11_WBF_infernece.ipynb
β”œβ”€β”€ data_split.py
β”œβ”€β”€ eda.ipynb
β”œβ”€β”€ visualization_data.ipynb
└── wbf.py

πŸ“°Dataset Structure


β”œβ”€β”€ dataset
    β”œβ”€β”€ train.json
    β”œβ”€β”€ test.json
    β”œβ”€β”€ train
    └── test

πŸ“°Models

πŸ“°Backbones

πŸ“°Experiments

image

train_batch9921_41_46441c1a381ad986227e

Exp mAP
Yolov11(5), RT-DETR(5), CO-DINO(2) 0.6760
Co-dino_r50(2), Co-dino_swin(5) 0.6590
Co-dino_swin(5),Co-dino_r50(2),RT-DETR(5) 0.6797
Co_dino_swin(5), RT-DETR(5), Yolov11(5) 0.6834