{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T13:54:44Z","timestamp":1791554084412,"version":"4.3.4"},"reference-count":46,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,8,12]],"date-time":"2025-08-12T00:00:00Z","timestamp":1754956800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>Image classification is a highly significant field in machine learning (ML), especially when applied to address longstanding and challenging issues in the biological sciences, such as specie recognition and biodiversity conservation. In this study, we present the development of a hybrid machine learning-based tool suitable for deployment on mobile devices. This tool is aimed at processing and classifying three-dimensional samples of endemic lizard species from the Amazon rainforest. The dataset used in our experiment was collected at the Museu Paraense Em\u00edlio Goeldi (MPEG), Bel\u00e9m-PA, Brazil, and comprises three species: (a) <jats:italic>Anolis fuscoauratus<\/jats:italic>; (b) <jats:italic>Hoplocercus spinosus<\/jats:italic>; and (c) <jats:italic>Polychrus marmoratus<\/jats:italic>. We compared the effectiveness of four artificial neural networks (ANN) for feature extraction: (a) MobileNet; (b) MobileNetV2; (c) MobileNetV3-Small; and (d) MobileNetV3-Large. Additionally, we evaluated five classical ML models for classifying the extracted patterns: (a) Support Vector Machine (SVM); (b) GaussianNB (GNB); (c) AdaBoost (ADB); (d) K-Nearest Neighbors (KNN); and (e) Random Forest (RF). The performance metrics of all classifiers were very close, we used the McNemar\u2019s test on each model\u2019s confusion matrix to evaluate and compare their statistical significance. Our best model was a combination of a 2.9 million parameters MobileNetV3-Small as the feature extractor, with a linear kernel-based SVM as the classifier, which achieved accuracy of 0.955, precision of 0.948, recall of 0.948, and f1-score of 0.948. The results indicated that the use of a small deep learning (DL) model, in combination with a classical ML algorithm, emerges as a viable technique for classifying three-dimensional representations of lizard species samples. Such an approach facilitates taxonomic identification work for professionals in the field and provides a tool adaptable for integration into mobile data recording equipment, such as smartphones, and benefiting from more morphological features extracted from three-dimensional samples instead of two-dimensional images.<\/jats:p>","DOI":"10.3389\/frai.2025.1524380","type":"journal-article","created":{"date-parts":[[2025,8,12]],"date-time":"2025-08-12T05:31:35Z","timestamp":1754976695000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["A mobile hybrid deep learning approach for classifying 3D-like representations of Amazonian lizards"],"prefix":"10.3389","volume":"8","author":[{"given":"Arthur Gonsales","family":"da Silva","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Roger Pinho","family":"de Oliveira","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Caio","family":"de Oliveira Bastos","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Elena Almeida","family":"de Carvalho","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bruno Duarte","family":"Gomes","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,8,12]]},"reference":[{"key":"ref1","article-title":"Tensor flow: large-scale machine learning on heterogeneous distributed systems","author":"Abadi","year":"2016","journal-title":"arXiv"},{"key":"ref2","first-page":"415","article-title":"A novel technique for monitoring highly cryptic lizard species in forests","volume":"4","author":"Bell","year":"2009","journal-title":"Herpetol. Conserv. Biol."},{"key":"ref3","doi-asserted-by":"publisher","first-page":"1526","DOI":"10.3390\/ani13091526","article-title":"Animal species recognition with deep convolutional neural networks from ecological camera trap images","volume":"13","author":"Binta Islam","year":"2023","journal-title":"Animals"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pntd.0010647","article-title":"An artificial intelligence model to identify snakes from across the world: opportunities and challenges for Global Health and herpetology","author":"Bolon","year":"2022","journal-title":"SSRN Electron. J."},{"key":"ref5","doi-asserted-by":"publisher","first-page":"187","DOI":"10.1007\/s11160-024-09901-y","article-title":"Well-intentioned initiatives hinder understanding biodiversity conservation: an essay on a recent deep-learning image classifier for Amazonian fishes","volume":"35","author":"Campos","year":"2024","journal-title":"Rev. Fish Biol. Fish."},{"key":"ref6","doi-asserted-by":"publisher","first-page":"135","DOI":"10.5194\/we-23-135-2023","article-title":"Revisiting the debate: documenting biodiversity in the age of digital and artificially generated images","volume":"23","author":"Campos","year":"2023","journal-title":"Web Ecol."},{"key":"ref7","doi-asserted-by":"publisher","first-page":"9453","DOI":"10.1002\/ece3.5410","article-title":"Wildlife surveillance using deep learning methods","volume":"9","author":"Chen","year":"2019","journal-title":"Ecol. Evol."},{"key":"ref8","doi-asserted-by":"publisher","first-page":"e34013","DOI":"10.3897\/BDJ.7.e34013","article-title":"Dataset from the snakes (Serpentes, reptiles) collection of the Museu Paraense Em\u00edlio Goeldi, Par\u00e1, Brazil","volume":"7","author":"Da Costa Prudente","year":"2019","journal-title":"Biodivers Data J."},{"key":"ref9","doi-asserted-by":"publisher","first-page":"582110","DOI":"10.3389\/frai.2021.582110","article-title":"Supervised learning computer vision benchmark for Snake species identification from photographs: implications for herpetology and Global Health","volume":"4","author":"Durso","year":"2021","journal-title":"Front Artif Intell."},{"key":"ref10","article-title":"A tutorial on Bayesian optimization","author":"Frazier","year":"2018","journal-title":"arXiv"},{"key":"ref11","author":"Gill","year":"2024"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1016\/j.ecoinf.2017.07.004","article-title":"Towards automatic wild animal monitoring: identification of animal species in camera-trap images using very deep convolutional neural networks","volume":"41","author":"Gomez Villa","year":"2017","journal-title":"Ecol. Inform."},{"key":"ref9001","first-page":"1490","volume-title":"Methods Ecol Evol.","author":"Gray","year":"2019"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"151482","DOI":"10.1109\/ACCESS.2019.2947701","article-title":"Combining multiple feature-ranking techniques and clustering of variables for feature selection","volume":"7","author":"Haq","year":"2019","journal-title":"IEEE Access"},{"key":"ref14","article-title":"Searching for MobileNetV3","author":"Howard","year":"2019","journal-title":"arXiv"},{"key":"ref15","article-title":"Mobile nets: efficient convolutional neural networks for mobile vision applications","author":"Howard","year":"2017","journal-title":"arXiv"},{"key":"ref16","author":"Islam","year":"2020"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"771607","DOI":"10.3389\/fmed.2021.771607","article-title":"An introduction to machine learning approaches for biomedical research","volume":"8","author":"Jovel","year":"2021","journal-title":"Front. Med."},{"key":"ref18","author":"Kornblith","year":"2019"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"273","DOI":"10.1016\/j.patcog.2016.12.020","article-title":"A survey on image-based insect classification","volume":"65","author":"Martineau","year":"2017","journal-title":"Pattern Recogn."},{"key":"ref20","doi-asserted-by":"publisher","first-page":"112322","DOI":"10.1016\/j.rse.2021.112322","article-title":"Tree species classification from airborne hyperspectral and LiDAR data using 3D convolutional neural networks","volume":"256","author":"M\u00e4yr\u00e4","year":"2021","journal-title":"Remote Sens. Environ."},{"key":"ref21","doi-asserted-by":"publisher","first-page":"8137","DOI":"10.1038\/s41598-019-44565-w","article-title":"Insights and approaches using deep learning to classify wildlife","volume":"9","author":"Miao","year":"2019","journal-title":"Sci. Rep."},{"key":"ref22","doi-asserted-by":"publisher","first-page":"110","DOI":"10.1016\/j.procs.2019.08.147","article-title":"Mobilenet convolutional neural networks and support vector machines for palmprint recognition","volume":"157","author":"Michele","year":"2019","journal-title":"Procedia Comput. Sci."},{"key":"ref23","first-page":"1","article-title":"Cryptic multicolored lizards in the Polychrus marmoratus group (Squamata: Sauria: Polychrotidae) and the status of Leiolepis auduboni Hallowell","volume":"11","author":"Murphy","year":"2017","journal-title":"Amphib. Reptile."},{"key":"ref24","doi-asserted-by":"crossref","first-page":"499","DOI":"10.1007\/s44150-021-00015-8","article-title":"Insights into performance fitness and error metrics for machine learning","volume":"3","author":"Naser","year":"2023","journal-title":"Archit. Struct. Constr."},{"key":"ref25","doi-asserted-by":"publisher","first-page":"1070","DOI":"10.3390\/rs12071070","article-title":"Tree species classification of drone hyperspectral and RGB imagery with deep learning convolutional neural networks","volume":"12","author":"Nezami","year":"2020","journal-title":"Remote Sens"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"150","DOI":"10.1111\/2041-210X.13504","article-title":"A deep active learning system for species identification and counting in camera trap images","volume":"12","author":"Norouzzadeh","year":"2021","journal-title":"Methods Ecol. Evol."},{"key":"ref27","first-page":"2825","article-title":"Scikit-learn: machine learning in python","volume":"9","author":"Pedregosa","year":"2008","journal-title":"J. Mach. Learn."},{"key":"ref28","doi-asserted-by":"publisher","first-page":"184","DOI":"10.1093\/zoolinnean\/zlac087","article-title":"Identification of morphologically cryptic species with computer vision models: wall lizards (Squamata: Lacertidae: Podarcis) as a case study","volume":"198","author":"Pinho","year":"2023","journal-title":"Zool. J. Linnean Soc."},{"key":"ref29","author":"Polonen","year":"2018"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"93","DOI":"10.1186\/1471-2148-13-93","article-title":"A phylogeny and revised classification of Squamata, including 4161 species of lizards and snakes","volume":"13","author":"Pyron","year":"2013","journal-title":"BMC Evol. Biol."},{"key":"ref31","author":"Raju","year":"2020"},{"key":"ref32","doi-asserted-by":"publisher","first-page":"e9987","DOI":"10.1002\/ece3.9987","article-title":"Application of a deep learning image classifier for identification of Amazonian fishes","volume":"13","author":"Robillard","year":"2023","journal-title":"Ecol. Evol."},{"key":"ref33","author":"Rossum","year":"1995"},{"key":"ref34","author":"Sandler","year":"2018"},{"key":"ref35","doi-asserted-by":"publisher","first-page":"1529","DOI":"10.1007\/s42452-019-1568-z","article-title":"Machine vision based local fish recognition","volume":"1","author":"Sharmin","year":"2019","journal-title":"SN Appl. Sci."},{"key":"ref36","doi-asserted-by":"crossref","first-page":"347","DOI":"10.1007\/978-981-19-3015-7_25","article-title":"Classification of animals using MobileNet with SVM classifier","volume-title":"Computational methods and data engineering","author":"Sowmya","year":"2023"},{"key":"ref37","first-page":"117","article-title":"Microornamentation of lizard scales: some variations and taxonomic correlations","volume":"31","author":"Stewart","year":"1075","journal-title":"Herpetologica"},{"key":"ref38","doi-asserted-by":"publisher","first-page":"31","DOI":"10.1016\/j.ympev.2008.10.002","article-title":"Phylogeny of hoplocercine lizards (Squamata: Iguania) with estimates of relative divergence times","volume":"50","author":"Torres-Carvajal","year":"2009","journal-title":"Mol. Phylogenet. Evol."},{"key":"ref39","doi-asserted-by":"publisher","first-page":"792","DOI":"10.1038\/s41467-022-27980-y","article-title":"Perspectives in machine learning for wildlife conservation","volume":"13","author":"Tuia","year":"2022","journal-title":"Nat. Commun."},{"key":"ref40","first-page":"2579","article-title":"Visualizing data using t-SNE","volume":"9","author":"van der Maaten","year":"2008","journal-title":"J. Mach. Learn."},{"key":"ref41","doi-asserted-by":"publisher","first-page":"142","DOI":"10.1139\/z02-230","article-title":"Life above ground: ecology of Anolis fuscoauratus in the Amazon rain forest, and comparisons with its nearest relatives","volume":"81","author":"Vitt","year":"2003","journal-title":"Can. J. Zool."},{"key":"ref42","doi-asserted-by":"crossref","first-page":"2216","DOI":"10.1111\/2041-210X.13075","article-title":"Machine learning for image based species identification","volume":"9","author":"W\u00e4ldchen","year":"2018","journal-title":"Methods Ecol. Evol."},{"key":"ref43","author":"Wang","year":"2017"},{"key":"ref44","doi-asserted-by":"publisher","first-page":"533","DOI":"10.1111\/1365-2656.12780","article-title":"A computer vision for animal ecology. Prugh L, editor","volume":"87","author":"Weinstein","year":"2018","journal-title":"J. Anim. Ecol."},{"key":"ref45","doi-asserted-by":"publisher","first-page":"109347","DOI":"10.1016\/j.patcog.2023.109347","article-title":"A comprehensive survey of image augmentation techniques for deep learning","volume":"137","author":"Xu","year":"2023","journal-title":"Pattern Recogn."}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1524380\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,8,12]],"date-time":"2025-08-12T05:31:37Z","timestamp":1754976697000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1524380\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8,12]]},"references-count":46,"alternative-id":["10.3389\/frai.2025.1524380"],"URL":"https:\/\/doi.org\/10.3389\/frai.2025.1524380","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,8,12]]},"article-number":"1524380"}}