Gender classification in Bangla language using deep learning-based voice analysis

Abstract

Gender classification based on voice analysis is one of the elemental tasks in speech and audio processing, with various applications such as speech recognition systems, voice assistants, call center analytics, etc. For speech synthesis, speaker identification, and human-computer interaction- gender recognition plays a vital role. Although extensive research on this topic has been done in various languages, any studies can hardly be found regarding gender classification in the Bangla language. Our research paper aims to recognize gender in the Bangla language using deep learning approaches and voice analysis. The core of our approach involves the use of CNN models (ResNet50, EfficientNetB0, InceptionV3, and DenseNet-121) for our data training. The Mel-Frequency Cepstral Coefficients (MFCC) and short-time Fourier transforms (STFT) were computed from audio recordings and used as input features to the neural network model. The system’s excellent accuracy rate demonstrates its potential for use in practical settings. By providing light on the application of deep learning techniques in the context of the Bangla language, this study advances the area of gender identification. 95% accuracy was achieved in the InspectionV3 and EfficientNetB0 models with the MFCC input.