在台灣，石斑魚為水產養殖漁業中具有高經濟價值的魚類之一。為了在有限的單位面積上提高生產量，漁民以高密度養殖方式飼養石斑魚。然而，高密度養殖容易導致魚隻疾病傳染及提高死亡的風險。除日常檢查外，若漁民能夠及早辨識出某水下魚隻的異常外觀或情況，並進一步採取隔離措施，將有助於降低其他魚隻的感染機會。二十世紀以來，影像辨識技術已漸漸地應用於各個領域，但是限於在水下拍攝的影像環境複雜不穩定，而且在養殖場收集的水下魚隻影像也不充足，因此影像辨識於水產養殖漁業的應用仍佔少數。
為了提高水產養殖的表現，本研究提出一個兩階段ImageNet預訓練深度學習模型的應用。在第一階段，預先訓練的模型可以在十種不同的魚類中有效地分類出異常石斑魚。在第二階段，將正確分類的異常石斑魚影像作為預訓練模型的測試數據，預先訓練的模型能夠對異常石斑魚的四種異常外觀進行分類。
在這項研究中收集了7,700張水下魚隻的影像，來源包含線上開放資源和台灣屏東縣的一處石斑魚養殖場。此影像資料庫分為十一個類別，包括九種常見的高經濟價值魚類和具有正常和異常外觀的石斑魚。
本研究使用了四種不同結構的ImageNet預訓練模型，並使用實際水下影像進行驗證。實驗結果顯示，在第二階段中，InceptionV3預訓練模型能夠正確針對水下石斑魚四種不同類型的異常外觀進行分類，其平均準確率可達98.25％。

In Taiwan, grouper represents one of the high-economic-value fish species in aquaculture fishery. To achieve more production in limited land, fishermen breed grouper in a high-density manner. However, high-density breeding can cause stronger contagion and a higher risk of infection. In addition to routine inspections, if the fishermen can identify the abnormal appearance or condition of the underwater fish early and take further isolation measures, it will help reduce the chance of infection of other fish.
Nowadays, deep learning techniques have already implemented in various fields, but there are still insufficient applications on aquaculture fishery not only because of the complex and unstable environment underwater but also inadequate underwater images collected from the fish farms.
To improve aquaculture performance, a two-phase ImageNet pre-trained deep learning model is proposed in this study which is able to classify 4 types of the abnormal appearance of grouper. In this research, 7,700 underwater fish images are collected and divided into 11 classes averagely, including 9 common high-economic-value fish species and grouper with the normal and abnormal appearance in Taiwan.
There are four ImageNet pre-trained models are implemented in this paper and validated with empirical image data. The experimental result reveals that InceptionV3 pre-trained model is able to classify the four different types of the abnormal appearance of grouper which can reach average 98.25% Accuracy in phase II task.

A.I.Wiki. (2019). Open Datasets. Retrieved from skymind.ai: https://skymind.ai/wiki/open-datasets
Amara, J., Bouaziz, B., & Algergawy, A. (2017). A Deep Learning-based Approach for Banana Leaf Diseases Classification. BTW (Workshops), 79-88.
Andrew, W., Greatwood, C., & Burghardt, T. (2017). Visual localisation and individual identification of Holstein Friesian cattle via deep learning. IEEE International Conference on Computer Vision, 2850-2859.
Armando, V. (2016, June 23). The revolution of depth. Retrieved from https://medium.com/@Lidinwise/the-revolution-of-depth-facf174924f5
Barbedo, J. G. (2019). Plant disease identification from individual lesions and spots using deep learning. Biosystems Engineering, 180, 96-107.
Bruno, D. R., & Osório, F. S. (2017). Image classification system based on deep learning applied to the recognition of traffic signs for intelligent robotic vehicle navigation purposes. In 2017 Latin American Robotics Symposium (LARS) and 2017 Brazilian Symposium on Robotics (SBR). (pp. pp. 1-6). Curitiba, Brazil: IEEE.
Chang, C.-C., Lin, S.-H., Huang, M.-Y., Lin, K.-J., Chang, C.-I., & Lai, S.-S. (2014). Prevalence Rates of Diseases Among Farmed Grouper. Journal of Taiwan Fisheries Research, 22(2), 35-44.
Chen, C. L., & Qiu, G. H. (2014). The long and bumpy journey: Taiwan׳ s aquaculture development and management. Marine Policy, 48, 152-161.
Chen, S. W., Shivakumar, S. S., Dcunha, S., Das, J., Okon, E., Qu, C., & ... Kumar, V. (2017). Counting apples and oranges with deep learning: A data-driven approach. IEEE Robotics and Automation Letters, 2(2), 781-788.
Chen, T., Hong, W., & Wang, T. (2014). 石斑魚關鍵生物技術開發現況與趨勢. 農業生技產業季刊, pp. 38, 14-19.
Chollet, F. (2018). Deep Learning mit Python und Keras: Das Praxis-Handbuch vom Entwickler der Keras-Bibliothek. MITP-Verlags GmbH & Co. KG.
Ciresan, D. C., Meier, U., Masci, J., & Gambardella, L. M. (2011). Flexible, high performance convolutional neural networks for image classification. 22nd International Joint Conference on Artificial Intelligence.
DataSchool.io. (2014, March 25). Simple guide to confusion matrix terminology. Retrieved from https://www.dataschool.io/simple-guide-to-confusion-matrix-terminology/
Deng, J., Dong, W., Socher, R., Li, L. J., Li, K., & Fei-Fei, L. (2009, June). Imagenet: A large-scale hierarchical image database. 2009 IEEE conference on computer vision and pattern recognition (pp. 248-255). IEEE.
Dwivedi, P. (2019, Jan 5). Understanding and Coding a ResNet in Keras. Retrieved from https://towardsdatascience.com/understanding-and-coding-a-resnet-in-keras-446d7ff84d33
Dwivedi, P. (2019, Jan 5). Understanding and Coding a ResNet in Keras. Retrieved from Towards data science: https://towardsdatascience.com/understanding-and-coding-a-resnet-in-keras-446d7ff84d33
Egmont-Petersen, M., de Ridder, D., & Handels, H. (2002). Image processing with neural networks—a review. Pattern recognition, 35(10), 2279-2301.
Everingham, M. V. (2010). The pascal visual object classes (voc) challenge. International journal of computer vision, 88(2), pp. 303-338.
Feng, Y. A. (2012). Study on the Mixed Culture of Litopenaeus vannamei and Epinephelus malabaricus. Journal of Anhui Agricultural Sciences, 36.
Fuentes, A., Y. S., Kim, S., & Park, D. (2017). A robust deep-learning-based detector for real-time tomato plant diseases and pests recognition. Sensors, 17(9), 2022.
Google. (2016, September 30). Introducing the Open Images Dataset. Retrieved from Google AI Blog: https://ai.googleblog.com/2016/09/introducing-open-images-dataset.html
Grinblat, G. L., Uzal, L. C., Larese, M. G., & Granitto, P. M. (2016). Deep learning for plant identification using vein morphological patterns. Computers and Electronics in Agriculture, 127, 418-424.
Gulshan, V., Peng, L., Coram, M., Stumpe, M. C., Wu, D., Narayanaswamy, A., & ... Kim, R. (2016). Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. Jama, 316(22), 2402-2410.
Haralick, R. M., & Shanmugam, K. (1973). Textural features for image classification. IEEE Transactions on systems, man, and cybernetics, (6), 610-621.
He, K. M. (2015). Deep Residual Learning. Santiago, Chile.
He, K. Z. (2016). Deep residual learning for image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778). Las Vegas: IEEE.
Hegde, R. B., Prasad, K., Hebbar, H., & Singh, B. M. (2019). Comparison of traditional image processing and deep learning approaches for classification of white blood cells in peripheral blood smear images. Biocybernetics and Biomedical Engineering, 39.
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017). Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 4700-47081). IEEE.
Hubel, D. H., & Wiesel, T. N. (1959). Receptive fields of single neurones in the cat's striate cortex. The Journal of physiology, 148(3), 574-591.
Kamal, U., Rafi, A. M., Hoque, R., Das, S., Abrar, A., & Hasan, M. (2018). Application of DenseNet in Camera Model Identification and Post-processing Detection. arXiv preprint arXiv:1809.00576.
Kamilaris, A., & Prenafeta-Boldu, F. X. (2018). Deep learning in agriculture: A survey. Computers and Electronics in Agriculture, 147, 70-90.
Kotsiantis, S. B., Zaharakis, I., & Pintelas, P. (2007). Supervised machine learning: A review of classification techniques. Emerging artificial intelligence applications in computer engineering,160, 3-24.
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 1097-1105.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. nature, 521(7553), 436.
Lin, T. Y. (2014). European conference on computer vision. Microsoft coco: Common objects in context (pp. 740-755). Springer, Cham.
Litjens, G., Kooi, T., Bejnordi, B. E., S. A., Ciompi, F., Ghafoorian, M., . . . Sánchez, C. I. (2017). A survey on deep learning in medical image analysis. Medical image analysis, 42, 60-88.
Liu, B., Zhang, Y., He, D., & Li, Y. (2017). Identification of apple leaf diseases based on deep convolutional neural networks. Symmetry, 10(1), 11.
Lu, Y., Yi, S., Zeng, N., Liu, Y., & Zhang, Y. (2017). Identification of rice diseases using deep convolutional neural networks. Neurocomputing, 267, 378-384.
Marcelino, P. (2018, October 23). Towards Data Science. Retrieved from From a deep learning perspective, the image classification problem can be solved through transfer learning: https://towardsdatascience.com/transfer-learning-from-pre-trained-models-f2393f124751
Messelodi, S., Modena, C. M., & Zanin, M. (2005). A computer vision system for the detection and classification of vehicles at urban road intersections. Pattern analysis and applications, 8(1-2), 17-31.
MissingLink.ai. (2018). Retrieved from The Complete Guide to Artificial Neural Networks: Concepts and Models: https://missinglink.ai/guides/neural-network-concepts/complete-guide-artificial-neural-networks/
Ng, H. W. (2015). Deep learning for emotion recognition on small datasets using transfer learning. 2015 ACM on international conference on multimodal interaction (pp. 443-449). ACM.
Oppenheim, D., & Shani, G. (2017). Potato disease classification using convolution neural networks. Advances in Animal Biosciences, 8(2), 244-249.
Pan, S. J., & Yang, Q. (2009). A survey on transfer learning. IEEE Transactions on knowledge and data engineering, 22(10), 1345-1359.
Powers, D. M. (2011). Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation. Journal of Machine Learning Technologies, 2(1), 37-63.
Ranjan, R., Khan, A. R., Parikh, C., Jain, R., Mahto, R. P., Pal, S., & Chakravarty, D. (2016). Classification and identification of surface defects in friction stir welding: An image processing approach. Journal of Manufacturing Processes, 22, 237-25.
Rawat, W., & Wang, Z. (2017). Deep convolutional neural networks for image classification: A comprehensive review. Neural computation, 29(9), 2352-2449.
Ripley, B. D., & Hjort, N. L. (1996). Pattern recognition and neural networks. Cambridge: Cambridge university press.
Russell, B. C. (2008). LabelMe: a database and web-based tool for image annotation. International journal of computer vision, 77(1-3), pp. 157-173.
Saha, S. (2018, August). Towards Data Science. Retrieved from A Comprehensive Guide to Convolutional Neural Networks — the ELI5 way: https://towardsdatascience.com/a-comprehensive-guide-to-convolutional-neural-networks-the-eli5-way-3bd2b1164a53
Salman, A., Jalal, A., Shafait, F., Mian, A., Shortis, M., Seager, J., & Harvey, E. (2016). Fish species classification in unconstrained underwater environments based on deep learning. Limnology and Oceanography: Methods, 14(9), 570-585.
Santoni, M. M., Sensuse, D. I., Arymurthy, A. M., & Fanany, M. I. (2015). Cattle race classification using gray level co-occurrence matrix convolutional neural networks. Procedia Computer Science, 59, 493-502.
Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural networks, 61, 85-117.
Shafait, F., Mian, A., Shortis, M., Ghanem, B., Culverhouse, P. F., Edgington, D., & Harvey, E. S. (2016). Fish identification from videos captured in uncontrolled underwater environments. ICES Journal of Marine Science, 73(10), 2737-2746.
Shorten, C. (2019, Jan 25). Introduction to ResNets. Retrieved from Towards Data Science: https://towardsdatascience.com/introduction-to-resnets-c0a830a288a4
Shrimali, V. (2019, June 3). PyTorch for Beginners: Image Classification using Pre-trained models. Retrieved from Learn OpenCV: https://www.learnopencv.com/pytorch-for-beginners-image-classification-using-pre-trained-models/
Siddiqui, S. A., Salman, A., Malik, M. I., Shafait, F., Mian, A., Shortis, M. R., & Browman, H. (2017). Automatic fish species classification in underwater videos: exploiting pre-trained deep neural network models to compensate for limited labelled data. Marine Science, 75(1), 374-389.
Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint, p. arXiv:1409.1556.
Sladojevic, S., Arsenovic, M., Anderla, A., & Culibrk, D. &. (2016). Deep neural networks based recognition of plant diseases by leaf image classification. Computational intelligence and neuroscience.
Standford, V. L. (2010, April 30). Retrieved from http://www.image-net.org/about-stats?fbclid=IwAR2abR22GgQM0F6WXID0PxVqVZ8IXi62qMAs_oge79grz48SEJhSzPgKAh8
Standford, V. L. (2015). ImageNet: Large Scale Visual Recongnition Challenge (ILSVRC). Retrieved from http://www.image-net.org/challenges/LSVRC/
Szegedy, C. I. (2017). Inception-v4, inception-resnet and the impact of residual connections on learning. In Thirty-First AAAI Conference on Artificial Intelligence. AAAI.
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., & ... Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition, (pp. 1-9).
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016). Rethinking the inception architecture for computer vision. In Proceedings of the IEEE conference on computer vision and pattern recognition, (pp. 2818-2826).
Taiwan Fisheries Agency, C. o. (2017). Retrieved from Taiwan's fisheries statistical yearbook: https://goo.gl/SZst7H
Tajbakhsh, N., Shin, J. Y., Gurudu, S. R., Hurst, R. T., Kendall, C. B., Gotway, M. B., & Liang, J. (2016). Convolutional neural networks for medical image analysis: Full training or fine tuning? IEEE transactions on medical imaging, 35(5), 1299-1312.
Thilakarathne, H. (2018 , August 12). NaadiSpeaks. Retrieved from Deep Learning Vs. Traditional Computer Vision: https://naadispeaks.wordpress.com/2018/08/12/deep-learning-vs-traditional-computer-vision/
Troell, M., Joyce, A., Chopin, T., Neori, A., Buschmann, A. H., & Fang, J. G. (2009). Ecological engineering in aquaculture—Potential for integrated multi-trophic aquaculture (IMTA) in marine offshore systems. Aquaculture, 297(1-4), 1-9.
Villa, A. G., Salazar, A., & Vargas, F. (2017). Towards automatic wild animal monitoring: Identification of animal species in camera-trap images using very deep convolutional neural networks. Ecological informatics, 41, 24-32.
Wang, I. K., Yeh, S. L., & Chou, I. J. (2013). 生態化的水產養殖. 自然保育季刊,(82), 12-20.
Wutzl, B., Leibnitz, K., Rattay, F., Kronbichler, M., & Murata, M. (2019). Genetic algorithms for feature selection when classifying severe chronic disorders of consciousness. PLoS One, 14(7), e0219683.
Zahavy, T., Magnani, A., Krishnan, A., & Mannor, S. (2016). Is a picture worth a thousand words? A Deep Multi-Modal Fusion Architecture for Product Classification in e-commerce. arXiv preprint, arXiv:1611.09534.
Zhao, J.-b. (2011). 99年度建置水產動物生產醫學石斑魚教育訓練課程. 高雄縣, 台灣: 高雄縣動物防疫所.