로봇학회논문지 제14권 제2호 (통권 제52호) (p.139-149)

강건한 CNN기반 수중 물체 인식을 위한 이미지 합성과 자동화된 Annotation Tool

Synthesizing Image and Automated Annotation Tool for CNN based Under Water Object Detection
키워드 :
Deep Learning,Data Annotation,Object Detection,3D CAD Model

목차

Abstract
1. 서 론
2. 선행 연구 조사
  2.1 CNN기반 Object Detection
  2.2 3D CAD 모델을 사용한 이미지 생성
  2.3 Auto Annotation Tool
3. Auto Annotation tool
  3.1 Sample 생성
4. 이미지 합성
  4.1 Rendering Image
  4.2 Cropping Image
  4.3 Overlay Background
  4.4 수중 환경을 반영한 이미지 합성
  4.5 Translating synthetic image to haze image
5. Object Detection
  5.1 청수로 채워진 수조에서의 실험(Non Haze Dataset)
  5.2 실제 수중 환경을 반영한 수조에서의 실험(Haze Dataset)
6. 결 론
References

초록

In this paper, we present auto-annotation tool and synthetic dataset using 3D CAD model for deep learning based object detection. To be used as training data for deep learning methods, class, segmentation, bounding-box, contour, and pose annotations of the object are needed. We propose an automated annotation tool and synthetic image generation. Our resulting synthetic dataset reflects occlusion between objects and applicable for both underwater and in-air environments. To verify our synthetic dataset, we use MASK R-CNN as a state-of-the-art method among object detection model using deep learning. For experiment, we make the experimental environment reflecting the actual underwater environment. We show that object detection model trained via our dataset show significantly accurate results and robustness for the underwater environment. Lastly, we verify that our synthetic dataset is suitable for deep learning model for the underwater environments.