Journal of the Korea Institute of Information and Communication Engineering (한국정보통신학회논문지)

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지
DOI QR코드

DOI QR Code

Application of CNN for Fish Species Classification

어종 분류를 위한 CNN의 적용

  • Park, Jin-Hyun (Dept. of Mechatronics Engineering, Kyeognam National Univ. of Science and Technology) ;
  • Hwang, Kwang-Bok (Dept. of Mechatronics Engineering, Kyeognam National Univ. of Science and Technology) ;
  • Park, Hee-Mun (British American Tobacco Korea) ;
  • Choi, Young-Kiu (Department of Electrical Engineering, Pusan National University)
  • Received : 2018.11.07
  • Accepted : 2018.12.10
  • Published : 2019.01.31
Download PDF
⟨ Previous Next ⟩

Abstract

In this study, before system development for the elimination of foreign fish species, we propose an algorithm to classify fish species by training fish images with CNN. The raw data for CNN learning were directly captured images for each species, Dataset 1 increases the number of images to improve the classification of fish species and Dataset 2 realizes images close to natural environment are constructed and used as training and test data. The classification performance of four CNNs are over 99.97% for dataset 1 and 99.5% for dataset 2, in particular, we confirm that the learned CNN using Data Set 2 has satisfactory performance for fish images similar to the natural environment. And among four CNNs, AlexNet achieves satisfactory performance, and this has also the shortest execution time and training time, we confirm that it is the most suitable structure to develop the system for the elimination of foreign fish species.

본 연구에서 외래어종 퇴치를 위한 시스템 개발에 앞서 물 안의 어류 이미지를 CNN으로 학습하여 어종을 분류하는 알고리즘을 제안하고자 한다. CNN 학습을 위한 원데이터(raw data)는 각 어종에 대해 직접 촬영한 영상을 사용하였으며, 어종 분류성능을 높이기 위해 영상 이미지의 개수를 늘린 데이터세트 1과 최대한 자연환경과 가까운 영상 이미지를 구현한 데이터세트 2를 구성하여 학습 및 테스트 데이터로 사용하였다. 4가지 CNN의 분류성능은 데이터세트 1에 대해 99.97%, 데이터세트 2에 대해 99.5% 이상을 나타내었으며, 특히 데이터세트 2를 사용하여 학습한 CNNs이 자연환경과 유사한 어류 이미지에 대해서도 만족할 만한 성능을 가짐을 확인하였다. 그리고 4가지 CNN 중 AlexNet이 성능에서도 만족스러운 결과를 도출하였으며, 수행시간과 학습시간 역시 가장 짧아 외래어종 퇴치를 위한 시스템 개발에 가장 적합한 구조임을 확인하였다.

Keywords

HOJBC0_2019_v23n1_39_f0001.png 이미지

Fig. 1 CNN system[2-5]

HOJBC0_2019_v23n1_39_f0002.png 이미지

Fig. 2 data augmentation

HOJBC0_2019_v23n1_39_f0003.png 이미지

Fig. 3 Images in the water and composite fish images(a) Images of green algae and muddy water(b) Composite fish images

HOJBC0_2019_v23n1_39_f0004.png 이미지

Fig. 4 Training Set and Training Progress (a) Training Set 1 of AlexNet (b) Training Progress of AlexNet

HOJBC0_2019_v23n1_39_f0005.png 이미지

Fig. 5 Classification Result using Dataset 2 of AlexNet

HOJBC0_2019_v23n1_39_f0006.png 이미지

Fig. 6 Correct Predicted Results

HOJBC0_2019_v23n1_39_f0007.png 이미지

Fig. 7 Non Correct Predicted Results

Table. 1 The Dataset 1, Dataset 2

HOJBC0_2019_v23n1_39_t0001.png 이미지

Table. 2 The Classification Performance of CNNs using Test Data in Dataset 1

HOJBC0_2019_v23n1_39_t0002.png 이미지

Table. 3 The Classification Performance of CNNs learned by Dataset 1 using Test Data in Dataset 2

HOJBC0_2019_v23n1_39_t0003.png 이미지

Table. 4 The Classification Performance of CNNs using Test Data in Dataset 2

HOJBC0_2019_v23n1_39_t0004.png 이미지

References

  1. Korea Institute for International Economic Policy. 10th Session of the Conference of the Parties to the Convention on Biological Diversity : Nagoya Protocol [Internet]. Available: http://www.kiep.go.kr/sub/view.do?bbsId=globalecono&nttId=185515.
  2. Y. Le Cun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-based learning applied to document recognition," Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998. https://doi.org/10.1109/5.726791
  3. Y. Bengio, "Learning deep architectures for AI," Foundations and Trends in Machine Learning, vol. 2, no. 1, pp. 1-127, Feb. 2009. https://doi.org/10.1561/2200000006
  4. G. E. Hinton, S. Osindero, and Y. Teh, "A fast learning algorithm for deep belief nets," Neural Computation, vol. 18, no. 7, pp. 1527-1554, Jul. 2006. https://doi.org/10.1162/neco.2006.18.7.1527
  5. S. I. Hassan, L. Dang, S. H. Im, K. B. Min, J. Y. Nam, and H. J. Moon, "Damage Detection and Classification System for Sewer Inspection using Convolutional Neural Networks based on Deep Learning," Journal of the Korea Institute of Information and Communication Engineering, vol. 22, no. 3, pp. 451-457, Mar. 2018. https://doi.org/10.6109/JKIICE.2018.22.3.451
  6. J. H. Kim, D. S. Choi, H. S. Lee, and J. W. Lee, "Target Classification of Active Sonar Returns based on Convolutional Neural Network," Journal of the Korea Institute of Information and Communication Engineering, vol. 21 no. 10, pp. 1909-1916, Oct. 2017. https://doi.org/10.6109/JKIICE.2017.21.10.1909
  7. G. Chen, P. Sun, and Y. Shang, "Automatic Fish Classification System Using Deep Learning," Tools with Artificial Intelligence(ICTAI), 2017 IEEE 29th International Conference on. IEEE, pp. 24-29, 2017.
  8. V. A. Sindagi, and V. M. Patel, "A Survey of Recent Advances in CNN-based Single Image Crowd Counting and Density Estimation," Pattern Recognition Letters, vol. 107, no. 1, pp. 3-16, May 2018. https://doi.org/10.1016/j.patrec.2017.07.007
  9. M. Sarigul, and M. Avci, "Comparison of Different Deep Structures for Fish Classification," International Journal of Computer Theory and Engineering, vol. 9, no. 5, Oct. 2017.
  10. H. Qin, X. Li, J. Liang, Y. Peng, and C. Zhang, "DeepFish: Accurate under water live fish recognition with a deep architecture," Neurocomputing, vol. 187 no. 26, pp. 49-58, Apr. 2016. https://doi.org/10.1016/j.neucom.2015.10.122
  11. A. Salman, A. Jalal, F. S., A. Mian, M. Shortis, J. Seager, and E. Harvey, "Fish species classification in unconstrained underwater environments based on deep learning," LIMNOLOGY and OCEANOGRAPHY: METHODS, Association for the Sciences of Limnology and Oceanography, vol. 14, no. 9, pp. 570-585, Sep. 2016. https://doi.org/10.1002/lom3.10113
  12. Stanford Vision Lab, Stanford University, Princeton University, Large Scale Visual Recognition Challenge, [Internet]. Available: www.image-net.org.
  13. I. K Choi, H. E. Ahn, and J. S. Yoo, "Facial Expression Classification Using Deep Convolutional Neural Network," Journal of Electrical Engineering & Technology, vol. 13, no. 1, pp. 485-492, Jan. 2018. https://doi.org/10.5370/JEET.2018.13.1.485
  14. A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," Advances in neural information processing systems 25(NIPS2012), pp. 1097-1105, 2012.
  15. K. Simonyan and A. Zisserman. (2015, May). Very deep convolutional networks for large-scale image recognition. 3rd International Conference on Learning Representations. [Internet]. Available: http://arxiv.org/abs/1409.1556.
  16. MathWorks, Pretrained VGG-16 convolutional neural network, [Internet]. Available: https://kr.mathworks.com/help/deeplearning/ref/vgg16.html.
  17. MathWorks, Pretrained VGG-19 convolutional neural network, [Internet]. Available: https://kr.mathworks.com/help/deeplearning/ref/vgg19.html.
  18. MathWorks, Pretrained GoogLet convolutional neural network, [Internet]. Available: https://kr.mathworks.com/help/deeplearning/ref/googlenet.html.
  19. Y. Bengio, "Deep Learning of Representations for Unsupervised and Transfer Learning," Proceedings of Machine Learning Research, Volume 27: Proceedings of ICML Workshop on Unsupervised and Transfer Learning, Washington:WA, pp. 17-37, Jul. 2012.
  20. J. Ba, and D. P. Kingma. (2015, May). Adam: A Method for Stochastic Optimization. 3rd International Conference on Learning Representations. [Internet]. Available: http://arxiv.org/abs/1412.6980.

Cited by

  1. 반려동물 모니터링을 위한 YOLO 기반의 이동식 시스템 설계 vol.24, pp.1, 2019, https://doi.org/10.6109/jkiice.2020.24.1.22
  2. 합성곱 신경망을 이용한 구글 어스에서의 녹지 비율 측정 vol.24, pp.3, 2020, https://doi.org/10.6109/jkiice.2020.24.3.349
  3. Deep Learning Applied to SEM Images for Supporting Marine Coralline Algae Classification vol.13, pp.12, 2019, https://doi.org/10.3390/d13120640