Samson Akinpelu

Dr. Samson Akinpelu

orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 July 2023 in IEEE Access
Verified
Journal Article
Open access
SASamson Akinpelu
Samson Akinpelu
Serestina ViririSerestina Viriri
Serestina Viriri
AAAdekanmi Adegun
Adekanmi Adegun
Speech Emotion Recognition (SER) system, which analyzes human utterances to determine a speaker’s emotion, has a growing impact on how people and machines interact. Recent growth in human-computer interaction and computational intelligence has drawn the attention of many researchers in Artificial Intelligence (AI) to deep learning because of its wider applicability to several fields, including computer vision, natural language processing, and affective computing, among others. Deep learning models do not need any form of manually created features because they can automatically extract the prospective features from the input data. Deep learning models, however, call for a lot of resources, high processing power, and hyper-parameter tuning, making them unsuitable for lightweight devices. In this study, we focused on developing an efficient lightweight model for speech emotion recognition with optimized parameters without compromising performance. Our proposed model integrates Random Forest and Multi-layer Perceptron(MLP) classifiers into the VGGNet framework for efficient speech emotion recognition. The proposed model was evaluated against other deep learning based methods (InceptionV3, ResNet, MobileNetV2, DenseNet) and it yielded low computational complexity with optimum performance. The experiment was carried out on three datasets of TESS, EMODB, and RAVDESS, and Mel Frequency Cepstral Coefficient(MFCC) features were extracted with 6-8 variants of emotions namely, Sad, Angry, Happy, Surprise, Neutral, Disgust, Fear, and Calm. Our model demonstrated high performance of 100%, 96%, and 86.25% accuracy on TESS, EMODB, and RAVDESS datasets respectively. This revealed that the proposed lightweight model achieved higher accuracy of recognition compared to the recent state-of-the-art model found in the literature.

18 August 2022 in Applied Sciences
Verified
Journal Article
Open access
PDF file
SASamson Akinpelu
Samson Akinpelu
Serestina ViririSerestina Viriri
Serestina Viriri
Speech Emotion Classification (SEC) relies heavily on the quality of feature extraction and selection from the speech signal. Improvement on this to enhance the classification of emotion had attracted significant attention from researchers. Many primitives and algorithmic solutions for efficient SEC with minimum cost have been proposed; however, the accuracy and performance of these methods have not yet attained a satisfactory point. In this work, we proposed a novel deep transfer learning approach with distinctive emotional rich feature selection techniques for speech emotion classification. We adopt mel-spectrogram extracted from speech signal as the input to our deep convolutional neural network for efficient feature extraction. We froze 19 layers of our pretrained convolutional neural network from re-training to increase efficiency and minimize computational cost. One flattened layer and two dense layers were used. A ReLu activation function was used at the last layer of our feature extraction segment. To prevent misclassification and reduce feature dimensionality, we employed the Neighborhood Component Analysis (NCA) feature selection algorithm for picking out the most relevant features before the actual classification of emotion. Support Vector Machine (SVM) and Multi-Layer Perceptron (MLP) classifiers were utilized at the topmost layer of our model. Two popular datasets for speech emotion classification tasks were used, which are: Berling Emotional Speech Database (EMO-DB), and Toronto English Speech Set (TESS), and a combination of EMO-DB with TESS was used in our experiment. We obtained a state-of-the-art result with an accuracy rate of 94.3%, 100% specificity on EMO-DB, and 97.2%, 99.80% on TESS datasets, respectively. The performance of our proposed method outperformed some recent work in SEC after assessment on the three datasets.

4 October 2024 in IEEE Access
Verified
Journal Article
Open access
SASamson Akinpelu
Samson Akinpelu
Serestina ViririSerestina Viriri
Serestina Viriri
The intricate landscape of speech emotion classification poses a captivating yet challenging realm due to emotions being fundamental to human communication. In recent years, deep learning frameworks have emerged as powerful tools, shedding light on the elusive domain of emotion recognition, revolutionizing human-computer interactions, and enhancing the emotional intelligence of artificial intelligence (AI). This survey embarks on an exploratory journey into the forefront of deep learning approaches dedicated to speech emotion classification. Deep learning has become the standard approach due to the scarcity of extensive speech corpora and the need for high accuracy at low computational cost. The reason lies in its potency to extract important emotional features from large or medium-sized spectrogram images. Deep learning has been applied to speech emotion classification by many researchers, leading to significant improvements in performance and accuracy. Modern deep learning methods designed for human auditory speech emotion classification are carefully examined in this work. A thorough examination of various deep learning framework designs used in emotion classification is provided, illuminating unique characteristics that capture essential features from speech signals for accurate emotion prediction. The research critically analyzes selected deep models using well-established emotion corpora, highlighting their effectiveness. This research analyses typical performance evaluation metrics used to evaluate speech emotion classification models. With this review, we hope to offer a comprehensive overview of the state-of-the-art, potential directions for further investigation, and developing approaches that further the field of speech emotion classification with deep learning frameworks.

28 April 2025 in ParadigmPlus
Verified
Journal Article
Open access
PDF file
SASamson Akinpelu
Samson Akinpelu
OOOdunayo Emmanuel Olasoji
Odunayo Emmanuel Olasoji
AAAkindayo Akindolani
Akindayo Akindolani
KAKorede Israel Adeyanju
Korede Israel Adeyanju
Sunday Adeola AJAGBESunday Adeola AJAGBE
Sunday Adeola AJAGBE
GTGbadegesin Adetayo Taiwo
Gbadegesin Adetayo Taiwo
Small-scale farmers use cassava as an essential crop for food and nutrition security, owing to its capacity to flourish under adverse situations. In numerous African countries, it serves as a significant source of carbohydrates. Leaf diseases can sometimes damage cassava crops, constraining overall output and farmers' revenue. The ongoing research on cassava disease is fraught with challenges, including a low detection rate, extended processing time, and inadequate precision. This research employs deep transfer learning with Visual Geometry Group (VGG16) models for the diagnosis of Cassava leaf diseases. An experimental study is conducted on a dataset of 5,656 images of cassava, categorized into four distinct disease classifications. Two of the most sophisticated predictions were generated: one identified the healthy leaf, while the other detected the diseases present on the unhealthy leaf. Our suggested deep transfer learning model attains a promising accuracy of 88% and an F1-score of 82% on the public plant disease dataset from the Kaggle repository, achieved through effective hyperparameter fine-tuning. The results of this study strongly advocate for additional research and practical application of deep learning models in plant disease diagnostics.

15 July 2025 in Cognitive Computation
Verified
Journal Article
Open access
PDF file
SASamson Akinpelu
Samson Akinpelu
Serestina ViririSerestina Viriri
Serestina Viriri
MYMuhammad Haroon Yousaf
Muhammad Haroon Yousaf
Emotion recognition from human speech occupies a significant position in Human-Computer Interaction, especially with the recent advancements in Artificial Intelligence and Robotic computing. As the level of interactivity of man–machine increases, intuitive responses that are emotionally based have attracted a lot of research into emotion recognition from speech signals. However, with various machine learning models littering the literature, cross-language efficient speech emotion recognition with extracted features inherent in speech signals with state-of-the-art deep learning techniques, is still posing a serious challenge. In this paper, we proposed a deep learning transformer network based on a shift window for speech emotion recognition using speech corpus from two different languages. Shift Window Transformer (SWT) is based on a hierarchical transformer architecture designed for natural language tasks and has recently become a novel model in computer vision and image processing tasks. The input feature to the model, Mel spectrogram, is extracted from two public speech datasets: Toronto English Emotion Speech (TEES) and EMOVO. Our proposed transformer model achieved a promising result of 98.3%, 64%, and 66% recognition accuracy on TESS, EMOVO, and TESS_EMOVO (hybrid bi-lingual) datasets, respectively, after extensive experiments and parameter optimization. Our performance evaluation revealed that the proposed model yielded an improved result in the recognition of six different emotions from human auditory speech compared to others found in the literature. The study explores the performance of the SWT architecture on cross-language speech emotion recognition and informs future robust and adaptive model development.

13 March 2025 in SN Computer Science
Verified
Journal Article
Open access
PDF file
CDComfort Yetunde Daramola
Comfort Yetunde Daramola
SASamson Akinpelu
Samson Akinpelu
EAEzekiel Olajuwon Akinyemi
Ezekiel Olajuwon Akinyemi
Sunday Adeola AJAGBESunday Adeola AJAGBE
Sunday Adeola AJAGBE
GAGabriel Akinyemi Akinpelu
Gabriel Akinyemi Akinpelu
MAMatthew Olusegun Adigun
Matthew Olusegun Adigun
The successful development of an effective machine learning model for detecting malicious queries is a challenging task that requires domain expertise. Malicious queries are complex and constantly evolving, and their detection is further complicated by variations in database server architectures. To tackle this challenge, domain experts with a deep understanding of attack vectors, techniques, and trends are crucial. Their insights into the patterns, behaviour, and characteristics of malicious queries provide invaluable guidance for designing appropriate features and training the model. This study aims to leverage this domain expertise to propose a robust machine-learning model capable of accurately identifying and categorizing malicious queries. By incorporating expert knowledge and using advanced techniques like machine learning, the model can be optimized to address the dynamic nature of these threats, leading to improved accuracy in real-world scenarios. After the extensive experiment, using 88,213 malicious and normal queries from GitHub and other sources, the Random Forest model yielded the highest accuracy of 98.4% and 97% sensitivity compared to other machine learning models experimented in the study. The performance comparison of the study with existing works indicates that the machine learning model achieved promising results.