Earticle

초록

영어

In the agricultural field, the application of deep convolutional neural networks is increasing. Especially, in the task such as harvesting, instance-level segmentation is required to target fruits. Even though a large amount of data is required to train instance segmentation, it is not easy to obtain sufficient dataset for tomatoes. Therefore, synthetic images are generated through data augmentation for tomato instance segmentation. The training is performed through small number of real images and augmented images. As a result of training from real images, the best accuracy is 73.47%. Based on the synthetic data augmentation, the best accuracy is 89.87% with the generation of maximum 3 foreground objects per an image. We also show that the results of tomato detection and instance segmentation qualitatively.

목차

Abstract
I. INTRODUCTION
II. TRAINING IMAGE DATA
A. Real image data
B. Synthetic datasets with data augmentation
III. MASK-RCNN MODEL TRAINING
A. The framework of Mask-RCNN
B. Training through actual images and composition images
IV. TRAINING RESULT
A. Actual image data training result
B. Synthetic image data training result
V. CONCLUSION
ACKNOWLEDMENTS
REFERENCES

키워드

저자

• Min-Ho Jang [ Department of Biosystem Engineering Chungbuk National University ]
• Chang-Seop Shin [ Department of Biosystem Engineering Chungbuk National University ]
• Chung-Gi Ban [ Department of Intelligent Systems & Robotics Chungbuk National University ]
• Youngbae Hwang [ Department of Intelligent Systems & Robotics Chungbuk National University ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  한국차세대컴퓨팅학회 학술대회

• 간기

  반년간

• 수록기간

  2021~2025

• 십진분류

  KDC 566 DDC 004