Animal Species Classification Using Fine-Tuned ResNet152V2: A Deep Learning Approach for Biodiversity Conservation

Especially in challenging fields like animal species identification, deep learning methods CNN have greatly advanced image categorization. Ten different animal species Butterfly, Cat, chicken, cow, dog, elephant, Horse, sheep, spider, and squirrel, are classified using ResNet 152V2, a deep CNN architecture, in this work. The work uses class weighting, data augmentation, and transfer learning to handle class imbalance so enhancing the the model ability to generalise. A validation accuracy of 88% results from fine-tuning the ResNet 152V2 model to maximize accuracy and economy. With an eye toward minority groups like elephants, the classification performance was fully assessed utilizing precision, recall, and F1-scores. Consistent progress without overfitting indicated the robustness of the model by means of the examination of training and validation loss and accuracy. By proving the possibilities of deep CNNs in uses like wildlife monitoring, biodiversity preservation, sustainable development, ecosystem preservation, and environmental resilience, this research supports long-term conservation efforts and responsible use of natural resources so contributing to the field of image classification.