<![CDATA[Indonesia is known for its abundant natural resources, including its diverse bird fauna. The identification and classification of bird species is essential in maintaining biodiversity as well as for practical habitat management. Therefore, an efficient and accurate approach is needed to identify bird species. This study uses a deep learning approach to test and compare the MobileNetV2 architecture with architectures used in previous studies in recognizing bird species. We use a transfer learning approach that utilizes existing knowledge from pre-trained models and combines it with a Convolutional Neural Network (CNN) algorithm to detect and classify birds based on images with a total image data of 95376. Experimental results show that by using the MobileNetV2 architecture, we achieved an accuracy of 96.4% with a loss value of 0.241. Compared with the architecture used in previous research, our results show a significant improvement in accuracy and efficiency. The time taken to perform the classification at each step is about 646 ms. This study shows that using MobileNetV2 architecture in the transfer learning approach with CNN effectively performs bird species classification.]]>
Copyrights © 2023
