- 교통사고 영상 수집
- ffmpeg를 통한 교통사고 사진 수집
- YOLO-mark를 통한 labeling
- yolov3.cfg파일 수정 (classes=1, filters=64)
- 교통사고 모델 학습
- 임계값을 0.7로 두어 학습된 사고와 70%이상의 일치율으르 보일 때 사고 감지
- test data를 통한 모델 테스트
-
💻 Computer specs used to train the model CPU : i5-7300HQ, GPU : GTX 1050 4GB
-
📒 Data used to train the model : https://www.youtube.com/watch?v=kVl-6-A9ZO4&list=PL8v-ndW9Fx1cLTxFtfd4HLc7k7NGF-xJa
-
📮 Contact us : sheewons@naver.com or u_rim@naver.com
-
🇰🇷 Department of Information and Communication Engineering, Hankuk University of Foreign Studies
