Uso de Redes Neuronales Convolucionales para el Reconocimiento Automático de Imágenes de Macroinvertebrados para el Biomonitoreo Participativo
Abstract
In Panama, there are community organizations that guarantee access to water for human consumption to more than 20% of the country's total population. For the sustainability of the water resource, it is essential to involve the communities in the process of monitoring the water quality. This can be achieved through the implementation of participatory biomonitoring using macroinvertebrates as indicators. In fact, it has been determined that the presence of different families of these organisms in ecosystems can be associated to different levels of their ecological quality. This work aims to develop a system capable of recognizing two families of macroinvertebrates through the use of images. The system is based on the use of algorithms of deep neural networks, with which we can achieve the learning of patterns. From a set of public images from the internet and biomonitoring carried out in the field, we train a convolutional neural network implemented in Tensorflow and Keras. These images belong to photographs of specimens of the families Calopterygidae and Heptageniidae. For this preliminary test, we report reliability percentages with values above 95%.
Keywords: image recognition, neural networks, convolutional neural networks, macroinvertebrates
References
[1] Brownlee, J. (2016). 8 Inspirational Applications of Deep Learning.
[2] Boyle, S., Somma, A., Codjia, J., Ure, J., Romanelli L., & Momo, F. (2008). Recognition of freshwater macroinvertebrate taxa by image analysis and artificial neural networks, In Proceedings of the International Symposium on Mathematical and Computational Biology, Campos do Jordäao, Brazil, 22-27 November 2008.
[3] Cireşan, D. C., Meier, U., Gambardella, L. M., & Schmidhuber, J. (2010). Deep, big, simple neural nets for handwritten digit recognition. Neural computation, 22(12), 3207-3220.2
[4] Coutinho, E., Weninger, F., Schuller, B. W., & Scherer, K. R. (2014, October). The Munich LSTM-RNN Approach to the MediaEval 2014” Emotion in Music’” Task. In MediaEval.
[5] Dietterich, T.G., Lytle, D.A., Mortensen, E.N., Moldenke, A.R., Paasch, R.K., Sarpola, M.J., & Shapiro, L.G. (2009). An Aquatic Insect Imaging System to Automate Insect Classification. Transactions of the American Society of Agricultural and Biological Engineers.
[6] Forbes. (2017, January). 10 Powerful Examples of Artificial Intelligence in Use Today. (R. Adams, Productor).
[7] Ioffe, S., & Szegedy, C. (2015, June). Batch normalization: Accelerating deep network training by reducing internal covariate shift. In International Conference on Machine Learning (pp. 448-456).
[8] Goodfellow, Ian, Bengio, Yoshua and Courville, Aaron. (2016, December). Deep Learning. MIT Press.
[9] Karpathy, A. (2015). The unreasonable effectiveness of recurrent neural networks. Andrej Karpathy blog.
[10] Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems (pp. 1097-1105).
[11] LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
[12] Oquab, M., Bottou, L., Laptev, I., & Sivic, J. (2014). Learning and transferring mid-level image representations using convolutional neural networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 1717-1724).
[13] Saldaña Fabela, M. d., López López, R., & Salcedo Sánchez, E. (2001). Utilización de un índice de diversidad para determinar la calidad del agua en sistemas lóticos. Ingeniería hidráulica en México, Abr-jun, pp. 57-66, 2001.
[14] 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).
[15] 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).
[16] Tirronen, V., Caponio, A., Haanpää, T., & Meissner, K. (2009). Multiple Order Gradient Feature for Macro-Invertebrate Identification Using Support Vector Machines. Adaptive and Natural Computing Algorithms. ICANNGA 2009. Lecture Notes in Computer Science, vol 5495. Springer, Berlin, Heidelberg.