DOI QR코드

DOI QR Code

전이학습에 방법에 따른 컨벌루션 신경망의 영상 분류 성능 비교

Comparison of Image Classification Performance in Convolutional Neural Network according to Transfer Learning

  • 투고 : 2018.08.28
  • 심사 : 2018.11.07
  • 발행 : 2018.12.31

초록

Core algorithm of deep learning Convolutional Neural Network(CNN) shows better performance than other machine learning algorithms. However, if there is not sufficient data, CNN can not achieve satisfactory performance even if the classifier is excellent. In this situation, it has been proven that the use of transfer learning can have a great effect. In this paper, we apply two transition learning methods(freezing, retraining) to three CNN models(ResNet-50, Inception-V3, DenseNet-121) and compare and analyze how the classification performance of CNN changes according to the methods. As a result of statistical significance test using various evaluation indicators, ResNet-50, Inception-V3, and DenseNet-121 differed by 1.18 times, 1.09 times, and 1.17 times, respectively. Based on this, we concluded that the retraining method may be more effective than the freezing method in case of transition learning in image classification problem.

키워드

MTMDCW_2018_v21n12_1387_f0001.png 이미지

Fig. 1. Weight freezing method of transfer learning.

MTMDCW_2018_v21n12_1387_f0002.png 이미지

Fig. 2. Weight retraining method of transfer learning.

MTMDCW_2018_v21n12_1387_f0003.png 이미지

Fig. 3. Example of a cropped insect image.

MTMDCW_2018_v21n12_1387_f0004.png 이미지

Fig. 4. Representative insect images used in the experiment.

MTMDCW_2018_v21n12_1387_f0005.png 이미지

Fig. 5. Comparison of accuracy and loss rates up to 100 epoch without early stopping in ResNet-50

MTMDCW_2018_v21n12_1387_f0006.png 이미지

Fig. 6. Comparison of Accuracy and loss rates up to 100 epoch without early stopping in Inception-V3.

MTMDCW_2018_v21n12_1387_f0007.png 이미지

Fig. 7. Comparison of Accuracy and loss rates up to 100 epoch without early stopping in DenseNet-121.

Table 1. Data set information

MTMDCW_2018_v21n12_1387_t0001.png 이미지

Table 2. Operating system and middleware information used in the experiment

MTMDCW_2018_v21n12_1387_t0002.png 이미지

Table 3. Comparison of the two transfer learning results in ResNet-50

MTMDCW_2018_v21n12_1387_t0003.png 이미지

Table 4. The precision, recall, and f-score of ResNet-50

MTMDCW_2018_v21n12_1387_t0004.png 이미지

Table 5. Comparison of the two transfer learning results in Inception-V3

MTMDCW_2018_v21n12_1387_t0005.png 이미지

Table 6. The precision, recall, and f-score of Inception-V3

MTMDCW_2018_v21n12_1387_t0006.png 이미지

Table 7. Comparison of the two transfer learning results in DenseNet-121

MTMDCW_2018_v21n12_1387_t0007.png 이미지

Table 8. The precision, recall, and f-score of DenseNet-121

MTMDCW_2018_v21n12_1387_t0008.png 이미지

참고문헌

  1. L. Yann, B. Yoshua, and H. Geoffrey, “Deep Leaning,” Nature, Vol. 521, No. 7553, pp. 436-444, 2015. https://doi.org/10.1038/nature14539
  2. Y. Lecun, L. Bottou, and Y. Bengio, “Gradient-based Learning Applied to Document Recognition,” Proceeding of The IEEE, Vol. 86, No. 11, pp. 2278-2324, 1998. https://doi.org/10.1109/5.726791
  3. K. Alex, S. Ilya, and H. Geoffrey, "ImageNet Classification with Deep Convolutional Neural Networks," Proceeding of Advances in Neural Information Processing System, pp. 1097-1105, 2012.
  4. J. Donahue, Y. Jia, O. Vinyals, J. Hoffman, N. Zhang, E. Tzeng, et al., "DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition," arXiv, arXiv: 1310.1531, 2013.
  5. A.S. Razavian, H. Azizpour, J. Sullivan, and S. Carlsson, "CNN Features Off-the-shelf: An Astounding Baseline for Recognition," Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 512-519, 2014.
  6. K. He, X. Zhang, S. Ren, and J. Sun, "Deep Residual Learning for Image Recognition," arXiv, arXiv:1512.03385, 2015.
  7. C. Szegedy, V. Vanhoucke, S. loffe, J. Shlens, and Z. Wojna, "Rethinking the Inception Architecture for Computer Vision," arXiv, arXiv:1512.00567, 2015.
  8. G. Huang, Z. Liu, L. v. d. Maaten, and K. Q. Weinberger, "Densely Connected Convolutional Networks," Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 2261-2269, 2017.
  9. M. Abadi, A. Agarwal, P. Barham, E. Brevdo, Z. Chen, C. Citro et al., "TensorFlow: Large-scale Machine Learning on Heterogeneous Systems," arXiv, arXiv:1603.04467, 2015.
  10. Keras, https://github.com/fchollet/keras (accessed Jun., 4, 2018).
  11. Y. Jeong, l. Ansari, J. Shim, and J. Lee, "A Car Plate Area Detection System Using Deep Convolution Neural Network," Journal of Korea Multimedia Society, Vol. 20, No. 8, pp. 1166-1174, 2017. https://doi.org/10.9717/KMMS.2017.20.8.1166
  12. J. Yosinski, J. Clune, Y. Bengio, and H. Lipson, "How Transferable Are Features in Deep Neural Networks?," arXiv, arXiv:1411.1792, 2014.
  13. K. Janocha and W.M. Czarnecki, "On Loss Functions for Deep Neural Networks in Classification," arXiv, arXiv:1702.05659, 2017.
  14. V. Nair and G. Hinton, "Rectified Linear Units Improve Restricted Boltzmann Machines," Proceeding of the 27th International Conference on Machine Learning, pp. 807-814, 2010.
  15. M. Lin, Q. Chen, and S. Yan, "Network In Network," arXiv, arXiv:1312.4400v3, 2014.
  16. T. Tieleman and G. Hinton, Rmsprop: Divide the Gradient by a Running Average of Its Recent Magnitude, Coursera: Neural Networks for Machine Learning Technical Report, 2012.
  17. I. Sutskever, J. Martens, G. Dahl, and G. Hinton, "On the Importance of Initialization and Momentum in Deep Learning," Proceeding of the 30th International Conference on Machine Learning, Vol. 28, pp. 1139-1147, 2013.
  18. X. Glorot and Y. Bengio, "Understanding the Difficulty of Training Deep Feedforward Neural Networks," Proceeding of the International Conference on Artificial Intelligence and Statistics, pp. 249-256, 2010.
  19. K. He, X. Zhang, S. Ren, and J. Sun, "Delving Deep into Rectifiers: Surpassing Human-level Performance on Imagenet Classification," Proceeding of International Conference on Computer Vision, pp. 1026-1034, 2015.
  20. L. Prechelt, "Early Stopping-but When?," Neural Networks: Tricks of the Trade, pp. 53-67, 2012.
  21. Y. Bengio, "Practical Recommendations for Gradient-based Training of Deep Architectures," arXiv, arXiv:1206.5533, 2012.
  22. D. Jung, J. Son, and S. Kim, "Shot Category Detection Based on Object Detection Using Convolutional Neural Networks," Proceeding of International Conference on Advanced Communication Technology, pp. 36-39, 2018.
  23. S. Park, U. Park, and D. Kim, "Depth Image-based Object Segmentation Scheme for Simproving Human Action Recognition," Proceeding of International Conference on Electronics, Information, and Communication, pp. 1-3, 2018.

피인용 문헌

  1. 흉부 CT 영상에서 비소세포폐암 환자의 재발 예측을 위한 종양 내외부 영상 패치 기반 앙상블 학습 vol.24, pp.3, 2018, https://doi.org/10.9717/kmms.2020.24.3.373
  2. PCB 부품 검출을 위한 Knowledge Distillation 기반 Continual Learning vol.24, pp.7, 2018, https://doi.org/10.9717/kmms.2021.24.7.868