• Manasvi Ghanshyam Mali
• Amit Kumar Pandey
• Irfan Khan

This research advances bird species prediction by developing an enhanced Convolutional Neural Network (CNN) model based on MobileNetV2, tailored for lightweight, accurate classification in ecological applications. Leveraging TensorFlow and Keras, we employ transfer learning to adapt MobileNetV2, pre-trained on ImageNet, to classify 25 bird species from the CUB-200-2011 dataset, introducing a custom attention mechanism to boost feature discriminability. The dataset is pre-processed with advanced augmentation and normalization via image data generators, ensuring robustness across diverse conditions. MobileNetV2’s inverted residuals and linear bottlenecks optimize computational efficiency, achieving a test accuracy of 78.2%—outperforming baseline MobileNetV2 by 8%. Extensive evaluation on training, validation, and test sets, alongside field tests on new images, demonstrates the model’s real-world utility. Comparative analysis against ResNet-50 and EfficientNet-B0 highlights its superior efficiency-accuracy trade-off, with 300 million multiply-adds (MAdds). This work contributes a scalable, resource-efficient solution for biodiversity monitoring, with implications for mobile-based wildlife research and conservation.

"Bird Species Prediction Using Convolutional Neural Networks", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 3, page no.j34-j40, March-2025, Available :http://www.jetir.org/papers/JETIR2503905.pdf

"Bird Species Prediction Using Convolutional Neural Networks", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 3, page no. ppj34-j40, March-2025, Available at : http://www.jetir.org/papers/JETIR2503905.pdf