Abstract
Works requiring taxonomic knowledge face several challenges, such as arduous identification of many taxa and an insufficient number of taxonomists to identify a great deal of collected organisms. Machine learning tools, particularly convolutional neural networks (CNNs), are then welcome to automatically generate high-performance classifiers from available data. Supported by the image datasets available at the largest online database on ant biology, the AntWeb (www.antweb.org), we propose here an ensemble of CNNs to identify ant genera directly from the head, profile and dorsal perspectives of ant images. Transfer learning is also considered to improve the individual performance of the CNN classifiers. The performance achieved by the classifiers is diverse enough to promote a reduction in the overall classification error when they are combined in an ensemble, achieving an accuracy rate of over 80% on top-1 classification and an accuracy of over 90% on top-3 classification.
Highlights
Taxonomy is a cornerstone for biodiversity management, since information on species names and distributions is essential for scientific studies and environmental monitoring programs [1]
Aiming at developing a reliable machine learning algorithm for quickly identifying a large amount of specimens, we developed a Convolutional Neural Networks (CNNs) automated ant genera identification process that receive as input the head, profile and dorsum perspectives of ant images
In the Convolutional Neural Network Architecture section we describe the Type 1 CNN classifier and the architecture employed in this work, and in section CNN Training procedure we describe the train procedure applied to training the general and specific purpose CNNs
Summary
Taxonomy is a cornerstone for biodiversity management, since information on species names and distributions is essential for scientific studies and environmental monitoring programs [1]. Biodiversity research currently faces a series of obstacles which hinder the development of works founded on taxonomic knowledge [2, 3] Among these obstacles are the great number of specimens requiring identification and the concomitant low number of taxonomists available to perform this task [2, 4]. Aiming at developing a reliable machine learning algorithm for quickly identifying a large amount of specimens, we developed a CNN automated ant genera identification process that receive as input the head, profile and dorsum perspectives of ant images This dataset is a medium-size dataset; CNNs is capable of automatically detecting discriminant attributes—which is essential here for a high-performance ant genera identification—it needs a large size dataset, and over-fitting can be an issue. Such approach allows to deal with the possible issues arising from dataset sample size
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
