사이드 스캔 소나 영상에서 수중물체 자동 탐지를 위한 컨볼루션 신경망 기법 적용 = The application of convolutional neural networks for automatic detection of underwater object in side scan sonar images

다국어 초록 (Multilingual Abstract)

In this paper, we have studied how to search an underwater object by learning the image generated by the side scan sonar in the convolution neural network. In the method of human side analysis of the side scan image or the image, the convolution neural network algorithm can enhance the efficiency of the analysis. The image data of the side scan sonar used in the experiment is the public data of NSWC (Naval Surface Warfare Center) and consists of four kinds of synthetic underwater objects. The convolutional neural network algorithm is based on Faster R-CNN (Region based Convolutional Neural Networks) learning based on region of interest and the details of the neural network are self-organized to fit the data we have. The results of the study were compared with a precision-recall curve, and we investigated the applicability of underwater object detection in convolution neural networks by examining the effect of change of region of interest assigned to sonar image data on detection performance.

참고문헌 (Reference)

1 Seong-Ryul Kim, "The Reason Why to Use Acoustic Waves on the Sea-Bottom Survey" Korean Society of Marine Engineers 32 (32): 481-489, 2008

2 P. Blondel, "The Handbook of Side Scan Sonar" Springer Science & Business Media 23-, 2010

3 J. P. Fish, "Sound Underwater Images: A Guide to the Generation and Interpretation of Side Scan Sonar Data" Lower Cape Publishing Co. 11-47, 1990

4 L. Wan, "Regularization of neural networks using dropconnect" 1058-1066, 2013

5 D. Ciresan, "Multi-column deep neural networks for image classification" 3642-3649, 2012

6 A. Krizhevsky, "Imagenet classification with deep convolutional neural networks" 1097-1105, 2012

7 O. Russakovsky, "Image net large scale visual recognition challenge" 115 : 211-252, 2015

8 S. Ren, "Faster r-cnn: Towards real-time object detection with region proposal networks" 39 : 1137-1149, 2017

9 N. Srivastava, "Dropout: a simple way to prevent neural networks from overfitting" 15 : 1929-1958, 2014

10 K. He, "Delving deep into rectifiers: Surpassing human-level performance on imagenet classification" 1026-1034, 2015

11 Y. Sun, "Deep learning face representation by joint identification-verification" 1988-1996, 2014

12 I. Goodfellow, "Deep Learning" The MIT Press 199-216, 2016

13 S. Reed, "Automated approach to classification of mine-like objects in sidescan sonar using highlight and shadow information" 151 : 48-56, 2004

14 E. Dura, "Active learning for detection of mine-like objects in side-scan sonar imagery" 30 : 360-371, 2005

15 C. Goutte, "A probabilistic interpretation of precision, recall and F-score, with implication for evaluation" 5 : 345-359, 2005