Related Experiment Videos
A spectrogram and local feature-assisted convolutional neural network for Amharic speech emotion identification
Yeshambel Asmare Mengist1, Abrham Debasu Mengistu2, Mulatu Yirga Beyene3
1Department of Computer Science, Debark University, Debark, Ethiopia. yeshambeldebark2012@gmail.com.
Scientific Reports
|May 27, 2026
Summary
This study introduces a Convolutional Neural Network (CNN) model for Amharic Speech Emotion Recognition (SER). The CNN approach achieved 90% accuracy, outperforming other methods for low-resource languages.
Area of Science:
- Computational Linguistics
- Artificial Intelligence
- Speech Processing
Background:
- Speech Emotion Recognition (SER) is crucial for human-computer and human-human interaction.
- Low-resource languages like Amharic face challenges in SER due to limited datasets and linguistic diversity.
Purpose of the Study:
- To develop an efficient Amharic Speech Emotion Recognition system.
- To address the scarcity of data and linguistic diversity in low-resource SER.
Main Methods:
- A Convolutional Neural Network (CNN) architecture was employed.
- Combined spectrogram and local acoustic features (MFCCs, chroma, zero-crossing rate, energy, pitch) were utilized.
- Advanced preprocessing techniques including spectral subtraction and wavelet denoising were applied to a dataset of 1650 Amharic speech samples.
Main Results:
- The proposed CNN model achieved a classification accuracy of 90%.
- This significantly outperformed Recurrent Neural Network (RNN) based models (LSTM: 58.48%, BiLSTM: 63.33%, GRU: 40%) and single-feature models (local acoustic: 73%, spectrogram: 79%).
Conclusions:
- Integrating spectrogram and local acoustic features within a CNN architecture enhances accuracy and efficiency for Amharic SER.
- This approach sets a benchmark for future research in low-resource Speech Emotion Recognition.
Related Concept Videos
01:20Labeling Emotion
Emotional labeling is a cognitive process that involves identifying and naming one's emotions, such as anger, fear, happiness, or sadness. It allows individuals to recognize and express their internal emotional states, a critical aspect of emotional regulation and communication. Labeling emotions requires more than mere recognition; it also involves drawing upon memory and contextual cues to understand the current situation and apply a corresponding emotional label. For instance, feeling...
01:30Classification of Signals
In signal processing, signals are classified based on various characteristics: continuous-time versus discrete-time, periodic versus aperiodic, analog versus digital, and causal versus noncausal. Each category highlights distinct properties crucial for understanding and manipulating signals.
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...
A continuous-time signal holds a value at every instant in time, representing information seamlessly. In contrast, a discrete-time signal holds values only at specific moments, often denoted as x(n), where...