Determination of the best image recognition algorithm for South American camelids using machine learning, Apurímac, 2023
Keywords:
camelids, neural network, recognition, vicunaAbstract
Alpacas and vicuñas belong to the South American camelid family; the alpaca is a domestic animal, while the vicuña is a wild animal that generally lives at altitudes above 3,000 meters above sea level. The problem is that in the case of vicuñas it is difficult to recognize and count them because they are wild animals and because they flee from people and any other foreign object that is not part of their environment; the other problem is that there is no database of images or photos of vicuñas to carry out an image recognition study. This work consisted of collecting 146 vicuña images, of which 95 were considered for training, 41 for validation and 10 for testing; the images were collected in a village center in the district of Cotaruse, province of Aymaraes in the Apurimac region of Peru. In the case of alpacas, the images were obtained from the Kaggle repository and 142 images were used, of which 95 were considered for training, 41 for validation and 10 for testing. The results show that the Mask-RCNN algorithm or model obtains a value for accuracy of 1.0 for vicuñas and 0.95 for alpacas; these values are the most efficient in relation to those found by Yolo V8 and SSMD.
Downloads
References
H. D. Yacobaccio, “The domestication of South American camelids: a review,” Anim. Front., vol. 11, no. 3, pp. 43–51, 2021, https://doi.org/10.1093/af/vfaa065
J. E. Quispe Coaquira, T. Herrera Mamani, E. Apaza Zúñiga, L. Clavetea Quisca, and Z. Maquera Marón, “Características tecnológicas de la fibra de vicuñas en semicautiverio de la Multicomunal Picotani-Región Puno,” Rev. Investig. Vet. del Perú, vol. 29, no. 2, pp. 522–532, 2018, https://doi.org/10.15381/rivep.v29i2.14475
G. Rodríguez García, “Animales...? En peligro de extinción o en peligro de que los extingamos?,” Ius Prax., vol. 14, no. 1, pp. 301–315, 2008, doi: http://dx.doi.org/10.4067/S0718-00122008000100011.
J. Laker, J. Baldo, Y. Arzamendia, and H. D. Yacobaccio, “La vicuña en los Andes,” Investig. Conserv. y manejo vicuñas, pp. 37–50, 2006, [Online]. Available: https://www.vicam.org.ar/publi/VicuniaAndes.pdf
R. Hoffmann, K. Otte, and C. Ponce, El manejo de la vicuña silvestre Tomo I, vol. I. Alemania: Sociedad Alemana de Cooperación Técnica- GTZ, 1983.
G. Gallardo, A. Nuñez, and L. F. Pacheco, “Transectos lineales como opción para estimar abundancia de vicuñas (Vicugna vicugna): Estudio de caso en el Parque Nacional Sajama, Bolivia,” Ecol. en Boliv., vol. 45, pp. 64–72, 2010.
K. S. Christie, S. L. Gilbert, C. L. Brown, M. Hatfield, and L. Hanson, “Unmanned aircraft systems in wildlife research: Current and future applications of a transformative technology,” Front. Ecol. Environ., vol. 14, no. 5, pp. 241–251, 2016, doi: 10.1002/fee.1281, https://doi.org/10.1002/fee.1281
Kaggle, “Image Dataset,” 2023. https://www.kaggle.com/datasets/sid4sal/alpaca-dataset-small (accessed Jan. 02, 2023).
J. Hu, X. Wu, and M. Dai, “Estimating the population size of migrating Tibetan antelopes Pantholops hodgsonii with unmanned aerial vehicles,” Oryx, pp. 1–9, 2018, https://doi.org/10.1017/S0030605317001673
H. Nguyen et al., “Animal recognition and identification with deep convolutional neural networks for automated wildlife monitoring,” in 2017 IEEE International Conference on Data Science and Advanced Analytics (DSAA), 2017, pp. 40–49. https://doi.org/10.1109/DSAA.2017.31.
A. Gomez, G. Diez, A. Salazar, and A. Diaz, “Animal identification in low quality camera-trap images using very deep convolutional neural networks and confidence thresholds,” in International Symposium on Visual Computing, 2016, pp. 747–756. https://doi.org/10.1007/978-3-319-50835-1_67
J. Xiao, G. Liu, K. Wang, and Y. Si, “Cow identification in free-stall barns based on an improved Mask R-CNN and an SVM,” Comput. Electron. Agric., vol. 194, 2022, https://doi.org/10.1016/j.compag.2022.106738
S. Li et al., “Individual dairy cow identification based on lightweight convolutional neural network,” PLoS One, vol. 16, no. 11 November, 2021, https://doi.org/10.1371/journal.pone.0260510
M. Hussain, “YOLO-v1 to YOLO-v8, the Rise of YOLO and Its Complementary Nature toward Digital Manufacturing and Industrial Defect Detection,” Machines, vol. 11, no. 7, p. 677, 2023, https://doi.org/10.3390/machines11070677
K. He, G. Gkioxari, P. Dollár, and R. Girshick, “Mask r-cnn,” in Proceedings of the IEEE international conference on computer vision, 2017, pp. 2961–2969, https://doi.org/10.1109/ICCV.2017.322
S. A. Magalhães et al., “Evaluating the single-shot multibox detector and YOLO deep learning models for the detection of tomatoes in a greenhouse,” Sensors, vol. 21, no. 10, p. 3569, 2021, https://doi.org/10.3390/s21103569
CVAT, “CVAT-Open Data Annotation Platform,” 2023. https://www.cvat.ai/
Downloads
Published
Issue
Section
License
Copyright (c) 2023 C&T Riqchary Science and technology research magazine
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
