Capsule endoscopy detection DACON challenge

Overview

capsule_endoscopy_detection (DACON Challenge)

Overview

  • Yolov5, Yolor, mmdetection기반의 모델을 사용 (총 11개 모델 앙상블)
    • 모든 모델은 학습 시 Pretrained Weight을 yolov5, yolor, mmdetection 및 swin transformer github로부터 받아서 사용
    • 각 방식에 필요한 형태로 데이터의 format 변경
  • Train set과 Validation set을 나누어 진행
  • 총 11개의 결과를 앙상블
    • detectors_casacde_rcnn_resnet50_multiscale, retinanet_swin-l, retinanet_swin-l_multiscale, retinanet_swin-t, atss_swin-l_multiscale, faster_rcnn-swin-l_multiscale, yolor_tta_multiscale, yolov5x, yolov5x_tta, yolov5x_tta_multiscale
    • Weighted boxes fusion (WBF) 방식으로 앙상블 진행 (Iou threshold = 0.4)
    • 모델에 관한 보다 자세한 내용은 /all_steps 폴더 내에 STEP2로 시작하는 .sh 스크립트들에 적힌 주석을 참고해주세요!

환경(env) 세팅

  • 실험 환경: Ubuntu 18.04, Cuda 11.3, Anaconda3, Python 3.8
  1. git clone ( + 폴더 권한 설정)
git clone https://github.com/MAILAB-Yonsei/capsule_endoscopy_detection.git
chmod -R 777 capsule_endoscopy_detection
cd capsule_endoscopy_detection
  1. cbnet만 제외한 나머지에 대한 env 생성 (all_except_cbnet)
conda create -n all_except_cbnet python=3.8
conda activate all_except_cbnet
pytorch 설치 (ex. conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch)
pip install openmim
mim install mmdet
pip install -r requirements_all_except_cbnet.txt
conda deactivate
  1. cbnet에 대한 env 생성 (cbnet)
conda create -n cbnet python=3.8
conda activate cbnet
pytorch 설치 (ex. conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch)
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/{cu_version}/{torch_version}/index.html
     (ex. pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.10.0/index.html)
cd UniverseNet
pip install -r requirements/build.txt
pip install -v -e .  # or "python setup.py develop"
pip install instaboostfast
pip install git+https://github.com/cocodataset/panopticapi.git
pip install git+https://github.com/lvis-dataset/lvis-api.git
pip install albumentations>=0.3.2 --no-binary imgaug,albumentations
pip install pandas
pip install tqdm
pip install shapely
conda deactivate
cd ..

main code 실행

[각 STEP 별로 자세한 설명은 /all_steps 폴더 내의 각각의 .sh 파일에 적힌 주석을 참고해주세요!]

STEP0. data root path 지정

cd all_steps
gedit data_path.txt

data_path.txt 파일에 data의 절대 경로를 명시한다!!! (ex. /mnt/data)

STEP1. data preparation (약 20~30분 소요)

conda activate all_except_cbnet
bash STEP1_data_preparation.sh

STEP2. 각 모델을 학습시킨다. (pretrained 모델로 inference만 하고자 한다면 바로 STEP3로!)

  • cbnet만 제외한 나머지에 대한 Training
conda activate all_except_cbnet
bash STEP2_train_model1_atss_swin-l_ms.sh
bash STEP2_train_model2_detectors_cascade_rcnn_r50_ms.sh
bash STEP2_train_model3_faster_rcnn_swin-l_ms.sh
bash STEP2_train_model4_retinanet_swin-l.sh
bash STEP2_train_model5_retinanet_swin-l_ms.sh
bash STEP2_train_model6_retinanet_swin-t_ms.sh
bash STEP2_train_model7_yolor.sh
bash STEP2_train_model8_yolo5x.sh
  • cbnet에 대한 Training
conda activate cbnet
bash STEP2_train_model9_cbnet_faster_rcnn_swin-l_ms.sh

STEP3. 모든 모델에 대해 Inference를 진행한다. (모델 하나당 20~30분 소요)

  • STEP2.를 건너뛰고 pretrained 모델에 대해 test를 하는 경우 아래 과정을 수행한 뒤 STEP3.의 명령어를 실행:
    • 아래의 weight 파일 링크에서 받은 mmdetection/ckpts 폴더를 /mmdetection 폴더 아래에 위치시킨다.
    • 아래의 weight 파일 링크에서 받은 UniverseNet/ckpts 폴더를 /UniverseNet 폴더 아래에 위치시킨다.
    • 아래의 weight 파일 링크에서 받은 YOLO/ckpts 폴더를 /YOLO 폴더 아래에 위치시킨다.
    • weight 파일 링크: https://drive.google.com/drive/folders/151KJC3FTUsK5mfx4TtNbhiFuuvLIeGz-?usp=sharing
  • cbnet만 제외한 나머지에 대한 Inference
conda activate all_except_cbnet
bash STEP3_inference_all_except_cbnet.sh
  • cbnet에 대한 Inference
conda activate cbnet
bash STEP3_inference_cbnet.sh

SETP4. 모든 모델에 대해 앙상블을 진행한다.

conda activate all_except_cbnet
bash STEP4_ensemble.sh
  • 최종 파일은 가장 상위 디렉토리에 'final.csv'로 생성!!!

주의사항

모두 순서에 맞게 코드를 구성해놓았기 때문에 하나의 코드를 2번 실행하는 등의 경우 진행에 어려움이 있을 수 있습니다. 참고해주세요.

현재 코드는 validation은 진행하지 않게 주석처리했습니다. 원하시면 predict.py에서 validation 주석처리를 풀고 val_answer.csv 파일의 경로를 설정해주시면 됩니다.

(predict.py 파일 위치: /mmdetection/predict/main.py, /UniverseNet/predict/main.py)

Owner
MAILAB
Medical Artificial Intelligence Laboratory at Yonsei University, Republic of Korea
MAILAB
Jittor 64*64 implementation of StyleGAN

StyleGanJittor (Tsinghua university computer graphics course) Overview Jittor 64

Song Shengyu 3 Jan 20, 2022
Towards Understanding Quality Challenges of the Federated Learning: A First Look from the Lens of Robustness

FL Analysis This repository contains the code and results for the paper "Towards Understanding Quality Challenges of the Federated Learning: A First L

3 Oct 17, 2022
A no-BS, dead-simple training visualizer for tf-keras

A no-BS, dead-simple training visualizer for tf-keras TrainingDashboard Plot inter-epoch and intra-epoch loss and metrics within a jupyter notebook wi

Vibhu Agrawal 3 May 28, 2021
TipToiDog - Tip Toi Dog With Python

TipToiDog Was ist dieses Projekt? Meine 5-jährige Tochter spielt sehr gerne das

1 Feb 07, 2022
Stacked Hourglass Network with a Multi-level Attention Mechanism: Where to Look for Intervertebral Disc Labeling

⚠️ ‎‎‎ A more recent and actively-maintained version of this code is available in ivadomed Stacked Hourglass Network with a Multi-level Attention Mech

Reza Azad 14 Oct 24, 2022
tensorrt int8 量化yolov5 4.0 onnx模型

onnx模型转换为 int8 tensorrt引擎

123 Dec 28, 2022
Multistream CNN for Robust Acoustic Modeling

Multistream Convolutional Neural Network (CNN) A multistream CNN is a novel neural network architecture for robust acoustic modeling in speech recogni

ASAPP Research 37 Sep 21, 2022
Image Super-Resolution Using Very Deep Residual Channel Attention Networks

Image Super-Resolution Using Very Deep Residual Channel Attention Networks

kongdebug 14 Oct 14, 2022
Keras community contributions

keras-contrib : Keras community contributions Keras-contrib is deprecated. Use TensorFlow Addons. The future of Keras-contrib: We're migrating to tens

Keras 1.6k Dec 21, 2022
Implementation of STAM (Space Time Attention Model), a pure and simple attention model that reaches SOTA for video classification

STAM - Pytorch Implementation of STAM (Space Time Attention Model), yet another pure and simple SOTA attention model that bests all previous models in

Phil Wang 109 Dec 28, 2022
Civsim is a basic civilisation simulation and modelling system built in Python 3.8.

Civsim Introduction Civsim is a basic civilisation simulation and modelling system built in Python 3.8. It requires the following packages: perlin_noi

17 Aug 08, 2022
Add gui for YoloV5 using PyQt5

HEAD 更新2021.08.16 **添加图片和视频保存功能: 1.图片和视频按照当前系统时间进行命名 2.各自检测结果存放入output文件夹 3.摄像头检测的默认设备序号更改为0,减少调试报错 温馨提示: 1.项目放置在全英文路径下,防止项目报错 2.默认使用cpu进行检测,自

Ruihao Wang 65 Dec 27, 2022
Tutorial page of the Climate Hack, the greatest hackathon ever

Tutorial page of the Climate Hack, the greatest hackathon ever

UCL Artificial Intelligence Society 12 Jul 02, 2022
Keyword2Text This repository contains the code of the paper: "A Plug-and-Play Method for Controlled Text Generation"

Keyword2Text This repository contains the code of the paper: "A Plug-and-Play Method for Controlled Text Generation", if you find this useful and use

57 Dec 27, 2022
(CVPR2021) Kaleido-BERT: Vision-Language Pre-training on Fashion Domain

Kaleido-BERT: Vision-Language Pre-training on Fashion Domain Mingchen Zhuge*, Dehong Gao*, Deng-Ping Fan#, Linbo Jin, Ben Chen, Haoming Zhou, Minghui

250 Jan 08, 2023
Repository containing detailed experiments related to the paper "Memotion Analysis through the Lens of Joint Embedding".

Memotion Analysis Through The Lens Of Joint Embedding This repository contains the experiments conducted as described in the paper 'Memotion Analysis

Nethra Gunti 1 Mar 16, 2022
Official repository of the paper "A Variational Approximation for Analyzing the Dynamics of Panel Data". Mixed Effect Neural ODE. UAI 2021.

Official repository of the paper (UAI 2021) "A Variational Approximation for Analyzing the Dynamics of Panel Data", Mixed Effect Neural ODE. Panel dat

Jurijs Nazarovs 7 Nov 26, 2022
NumQMBasic - A mini-course offered to Undergrad physics students

The best way to use this material is by forking it by click the Fork button at the top, right corner. Then you will get your own copy to play with! Th

Raghu 35 Dec 05, 2022
Data loaders and abstractions for text and NLP

torchtext This repository consists of: torchtext.datasets: The raw text iterators for common NLP datasets torchtext.data: Some basic NLP building bloc

3.2k Jan 08, 2023
Third party Pytorch implement of Image Processing Transformer (Pre-Trained Image Processing Transformer arXiv:2012.00364v2)

ImageProcessingTransformer Third party Pytorch implement of Image Processing Transformer (Pre-Trained Image Processing Transformer arXiv:2012.00364v2)

61 Jan 01, 2023