https://hoya012.github.io/blog/yolov4/
https://medium.com/visionwizard
https://github.com/AlexeyAB/darknet
논문 링크
- You Only Look Once: Unified, Real-Time Object Detection
- YOLO9000: Better, Faster, Stronger
- YOLOv3: An Incremental Improvement
- YOLOv4: Optimal Speed and Accuracy of Object Detection
해설 영상 링크
- You Only Look Once: Unified, Real-Time Object Detection
- YOLO9000: Better, Faster, Stronger
- YOLOv3: An Incremental Improvement
- YOLOv4: Optimal Speed and Accuracy of Object Detection
필요조건
- Tensorflow 2.1.0
- tensorflow_addons 0.9.1 (required for mish activation)
설치방법
<< best requirment >>
- (darkNet)[https://github.com/AlexeyAB/darknet]
- CUDA 10.0
- cuDNN 7.4.3
- visual studio 2015
- opencv 4.1.0
- Downloading DarkNet
- (Alexeyab darknet)[https://github.com/AlexeyAB/darknet]
-
Copying Open files into DarkNet
-
Copying CuDNN files into DarkNet
-
Changing the CuDNN version in DarkNet
-
Compile YOLOv4 with update CUDA version
-
compiling DarkNet
-
Run Detction on Images
-
Detection on Videos
https://blog.naver.com/shuna11/221979528693
include / library 디렉토리 잡는 법
프로젝트 > 속성 > VC++ directories include directories
(cuda 설치 디렉토리)\include
(opencv\빌드\인스톨)\include
library directories
(cuda 설치 디렉토리)\lib\x64
(opencv\빌드\인스톨)\x64\vc14\lib
CUDA10.0.props & CUDA10.0.targets 잡는 법
line56
<Import Project="(cuda 설치 디렉토리)\extras\visual_studio_integration\MSBuildExtensions\CUDA10.0.props"/>
line308
<Import Project="(cuda 설치 디렉토리)\extras\visual_studio_integration\MSBuildExtensions\CUDA10.0.targets"/>