Speaker Recognition

Overview

Built a speaker recognition system using cepstrum-based features and CNN models.

Responsibilities

• Extracted MFCC/cepstrum features
• Designed CNN-based classification model
• Evaluated speaker identification accuracy

Approach

• Audio preprocessing
• Feature extraction (MFCC)
• CNN-based classification

Pipeline

flowchart LR
    A[Audio Input] --> B[Feature Extraction]
    B --> C[CNN Model]
    C --> D[Speaker Prediction]

Tech

PyTorch · Audio Processing · CNN

Impact

• Enabled accurate speaker identification
• Built foundation for voice-based AI systems