A Comparative Analysis of Clustering and Feature Extraction Methods for the Automated Construction of Bird Species Classification Datasets
Resumo
The identification of bird species enables the creation of machine learning models that can be employed for the non-invasive monitoring of bird populations. In this study, we present an advancement in the assisted automated creation of a training set for the classification of bird species, with a specific focus on species present in the Pantanal. Typically, this process is conducted manually, which is a highly time-consuming approach. In this phase, we propose comprehensive comparative testing to ascertain the optimal methodologies for feature extraction and clustering. Five clustering methods and four feature extraction models were subjected to testing. The results of our experiments demonstrate that the optimal method for the purpose of this work was hierarchical clustering, using BirdNET for feature extraction. This combination provided superior performance in classifying bird species for the assisted construction of training sets.
Referências
Chen, Y., Zhou, L., Bouguila, N., Wang, C., Chen, Y., and Du, J. Block-dbscan: Fast clustering for large scale data. Pattern Recognition vol. 109, pp. 107624, 2021.
Cheng, Y. Mean shift, mode seeking, and clustering. IEEE Transactions on Pattern Analysis and Machine Intelligence 17 (8): 790–799, 1995.
Choi, S., Baek, J., and Kim, D. Early diagnosis of combustion instability using statistical methods. Journal of the Korean Society of Combustion 27 (3): 1–7, 2022.
Cole, J. S., Michel, N. L., Emerson, S. A., and Siegel, R. B. Automated bird sound classifications of long-duration recordings produce occupancy model outputs similar to manually annotated data. Ornithological Applications 124 (2): duac003, 2022.
Dalton, D. T., Berger, V., Adams, V., Botha, J., Halloy, S., Kirchmeir, H., Sovinc, A., Steinbauer, K., Švara, V., and Jungmeier, M. A conceptual framework for biodiversity monitoring programs in conservation areas. Sustainability 15 (8): 6779, 2023.
Deng, D. Dbscan clustering algorithm based on density. In 2020 7th international forum on electrical engineering and automation (IFEEA). IEEE, Hefei, China, pp. 949–953, 2020.
JinHuaXu and HongLiu. Web user clustering analysis based on kmeans algorithm. In 2010 international conference on information, networking and automation (ICINA). Vol. 2. IEEE, Wuhan, China, pp. V2–6, 2010.
Kahl, S., Wood, C. M., Eibl, M., and Klinck, H. Birdnet: A deep learning solution for avian diversity monitoring. Ecological Informatics vol. 61, pp. 101236, 2021.
Kraskov, A., Stögbauer, H., Andrzejak, R. G., and Grassberger, P. Hierarchical clustering using mutual information. Europhysics Letters 70 (2): 278–284, 2005.
Krilašević, A., Mašetić, Z., and Kečo, D. Spotify playlist organization-mood-based cluster analysis. In 2024 23rd International Symposium INFOTEH-JAHORINA (INFOTEH). IEEE, Bosnia and Herzegovina, pp. 1–6, 2024.
Kumar, Y., Gupta, S., and Singh, W. A novel deep transfer learning models for recognition of birds sounds in different environment. Soft Computing 26 (3): 1003–1023, 2022.
Kvsn, R. R., Montgomery, J., Garg, S., and Charleston, M. Bioacoustics data analysis – a taxonomy, survey and open challenges. IEEE Access vol. 8, pp. 57684–57708, 2020.
Latham, P. E. and Roudi, Y. Mutual information. Scholarpedia 4 (1): 1658, 2009.
Liang, J., Nolasco, I., Ghani, B., Phan, H., Benetos, E., and Stowell, D. Mind the Domain Gap: a Systematic Analysis on Bioacoustic Sound Event Detection, 2024.
Liu, N., Xu, Z., Zeng, X.-J., and Ren, P. An agglomerative hierarchical clustering algorithm for linear ordinal rankings. Information Sciences vol. 557, pp. 170–193, 2021.
Liu, T. and Yuan, X. Paralinguistic and spectral feature extraction for speech emotion classification using machine learning techniques. EURASIP Journal on Audio, Speech, and Music Processing 2023 (1): 23, 2023.
Michaud, F., Sueur, J., Le Cesne, M., and Haupert, S. Unsupervised classification to improve the quality of a bird song recording dataset. Ecological Informatics vol. 74, pp. 101952, 2023.
Mirzal, A. Statistical analysis of microarray data clustering using nmf, spectral clustering, kmeans, and gmm. IEEE/ACM Transactions on Computational Biology and Bioinformatics 19 (2): 1173–1192, 2022.
Onishi, T., Yamauchi, A., Ogushi, A., Ishii, R., Fukayama, A., Nakamura, T., and Miyata, A. Modeling japanese praising behavior by analyzing audio and visual behaviors. Frontiers in Computer Science vol. 4, pp. 815128, 2022.
Pulatov, I., Oteniyazov, R., Makhmudov, F., and Cho, Y.-I. Enhancing speech emotion recognition using dual feature extraction encoders. Sensors 23 (14): 6640, 2023.
Stupariu, M.-S., Cushman, S. A., Pleşoianu, A.-I., Pătru-Stupariu, I., and Fuerst, C. Machine learning in landscape ecological analysis: a review of recent approaches. Landscape Ecology 37 (5): 1227–1250, 2022.
Terasaka, D., Martins, L., Santos, V., Ventura, T., Oliveira, A., and Pedroso, G. Audio segmentation to build bird training datasets. In Anais do XV Workshop de Computação Aplicada à Gestão do Meio Ambiente e Recursos Naturais. SBC, Porto Alegre, RS, Brasil, pp. 199–202, 2024.
van Osta, J. M., Dreis, B., Meyer, E., Grogan, L. F., and Castley, J. G. An active learning framework and assessment of inter-annotator agreement facilitate automated recogniser development for vocalisations of a rare species, the southern black-throated finch (poephila cincta cincta). Ecological Informatics vol. 77, pp. 102233, 2023.
Wu, K.-L. and Yang, M.-S. Mean shift-based clustering. Pattern Recognition 40 (11): 3035–3052, 2007.
Wu, S.-H., Ko, J. C.-J., Lin, R.-S., Chang-Yang, C.-H., and Chang, H.-W. Evaluating community-wide temporal sampling in passive acoustic monitoring: A comprehensive study of avian vocal patterns in subtropical montane forests. F1000Research vol. 12, pp. n.p, 2023.
Zhou, S., Chen, Z., Duan, R., and Song, W. Multi-exemplar affinity propagation clustering based on local density peak. Applied Intelligence 54 (3): 2915–2939, 2024.