The evolution of machine learning techniques in bird species identification: A Survey

Bird species identification through machine learning (ML) has emerged as a crucial tool in biodiversity conservation and ecological research. This study systematically reviews ML algorithms employed for bird species classification, emphasizing traditional approaches like k-Nearest Neighbors (KNN) and Support Vector Machines (SVM) alongside advanced deep learning techniques such as Feedforward Backpropagation Networks (FBN). Using the Xeno-canto dataset and MATLAB-based simulations, this research evaluates feature extraction methods, including Mel-Frequency Cepstral Coefficients (MFCCs), spectral, and timbre characteristics. Experimental results indicate that KNN and SVM achieved 100% accuracy with MFCC and spectral features, whereas FBN exhibited a slightly lower performance of 95-98%. The study highlights the importance of feature selection, model efficiency, and the impact of dataset variations. Additionally, classification challenges such as noise interference, dataset imbalance, and computational limitations are discussed. This review provides insights into the strengths and weaknesses of different ML techniques and suggests directions for enhancing automated bird species classification systems.