Muhammad Ahmed Abbasi

Dr. Muhammad Ahmed Abbasi

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
6 February 2024 in Signal, Image and Video Processing
Verified
Journal Article
Closed access
PDF file
MAMuhammad Ahmed Abbasi
Muhammad Ahmed Abbasi
HAHafza Faiza Abbasi
Hafza Faiza Abbasi
Muhammad Zulkifal AzizMuhammad Zulkifal Aziz
Muhammad Zulkifal Aziz
WHWaseem Haider
Waseem Haider
ZFZeming Fan
Zeming Fan
Xiaojun YuXiaojun Yu
Xiaojun Yu

1 October 2024 in Journal of Neural Engineering
Verified
Journal Article
Closed access
MAMuhammad Ahmed Abbasi
Muhammad Ahmed Abbasi
HAHafza Faiza Abbasi
Hafza Faiza Abbasi
Xiaojun YuXiaojun Yu
Xiaojun Yu
MAMuhammad Zulkifal Aziz
Muhammad Zulkifal Aziz
NYNicole Tye June Yih
Nicole Tye June Yih
ZFZeming Fan
Zeming Fan
Objective. Despite substantial advancements in Brain–Computer Interface (BCI), inherent limitations such as extensive training time and high sensitivity to noise largely hinder their rapid development. To address such issues, this paper proposes a novel extreme learning machine (ELM) based self-attention (E-SAT) mechanism to enhance subject-specific classification performances. Approach. Specifically, for E-SAT, ELM is employed both to improve self-attention module generalization ability for feature extraction and to optimize the model’s parameter initialization process. Meanwhile, the extracted features are also classified using ELM, and the end-to-end ELM based setup is used to evaluate E-SAT performance on different motor imagery (MI) EEG signals. Main results. Extensive experiments with different datasets, such as BCI Competition III Datasets IV-a, IV-b and BCI Competition IV Datasets 1, 2a, 2b, 3 are conducted to verify the effectiveness of the proposed E-SAT strategy. Results show that E-SAT outperforms several state-of-the-art and existing methods in subject-specific classification on all the datasets. An average classification accuracy of 99.8%, 99.1%, 98.9%, 75.8%, 90.8%, and 95.4% respectively is achieved for each datasets which demonstrate an improvement of 5%–6% compared to the existing methods. In addition, Kruskal Wallis test is performed to demonstrate the statistical significance of E-SAT and the results indicate significant difference with a 95% confidence level. Significance. The experimental results not only show outstanding performance of E-SAT in feature extraction, but also demonstrate that it helps achieve the best results among nine other robust classifiers. In addition, results in this study also demonstrate that E-SAT achieves exceptional performance in both binary and multi-class classification tasks, as well as for noisy and non-noisy datasets.

2 July 2025 in IEEE Access
Verified
Journal Article
Open access
HAHafza Faiza Abbasi
Hafza Faiza Abbasi
MAMuhammad Ahmed Abbasi
Muhammad Ahmed Abbasi
SJShen Jianbo
Shen Jianbo
XLXiang Liping
Xiang Liping
Xiaojun YuXiaojun Yu
Xiaojun Yu
Brain-computer interface (BCI) is inevitably a promising technology holding the potential to revolutionize the world with its wide range of applications. From healthcare to innovative computer gaming, integrating BCI for intelligent control has become an emergent scope. However, optimizing motor imagery (MI) classification in non-invasive BCI remains a significant challenge due to the poor quality of the acquired signal. In this paper, we propose a unified approach for MI classification by combining features from three diverse domains. Initially, the EEG data is preprocessed using bandpass filtering to extract the relevant EEG signals. Next, the preprocessed signal is fed simultaneously to three branches to extract three distinct categories of features from the signal. Specifically, spectral features are extracted using the fast-Fourier transform (FFT) and a spatial transformer is utilized to extract spatial features from the EEG data. Moreover, the third branch extracts temporal features using an encoder-decoder architecture. The features obtained using the three branches are concatenated together to obtain a comprehensive features set which is finally classified using extreme learning machine (ELM). Our proposed approach which uses a novel combination of features from three distinct domains is hereby named TriNet and is validated using two benchmark datasets BCI Competition IV-2a and BCI Competition IV-2b. The experimental results show an accuracy of 87.30% and 92.64% respectively on BCI IV-2a and BCI IV-2b datasets in subject-specific classification. Moreover, TriNet is also tested in subject-independent classification setup, and average classification accuracies of 63.92% and 78.60% are obtained on BCI IV-2a and 2b datasets respectively which is an improvement of 8 to 10% compared to the existing methods. The classification performance and computational cost comparisons demonstrate the superior performance of TriNet compared to the existing methods highlighting its potential to enhance MI classification in BCI.

11 August 2026 in Sensors
Verified
Journal Article
Open access
PDF file
MAMuhammad Ahmed Abbasi
Muhammad Ahmed Abbasi
HAHafza Faiza Abbasi
Hafza Faiza Abbasi
MAMuhammad Arsalan
Muhammad Arsalan
DKDanish Khan
Danish Khan
AAAndres Annuk
Andres Annuk
Xiaojun YuXiaojun Yu
Xiaojun Yu
Motor imagery (MI) classification serves as the backbone to brain–computer interfaces (BCIs) by strengthening the communication bridge between the human brain and external peripheral devices. The past two decades have witnessed unprecedented success in MI-BCIs, with applications not only in medical fields but also in several other domains, such as gaming and robotic control. Initially, MI classification primarily relied on classical signal processing techniques that were heavily impacted by signal variations; however, recent trends in deep learning (DL), specifically in sequence-oriented, attention-based, hybrid, and generative architectures such as recurrent neural networks (RNNs), variational autoencoders (VAEs), generative adversarial networks (GANs), and transformers have significantly improved the efficiency and robustness of MI classification. This study presents a comprehensive review of these sequence-oriented, attention-based, hybrid, and generative architectures including RNNs, VAEs, GANs, and transformers, comparing their robustness across various public MI datasets, highlighting their challenges, such as inter-subject variation, low signal-to-noise ratio (SNR), and the obstacles in real-time signal classification. We perform an in-depth analysis on the strengths and limitations of traditional models such as RNNs and LSTMs as well as emergent models such as VAEs and transformers, which have demonstrated superior performance in extracting the intricate patterns of the EEG data with low latency. Moreover, we critically examine the future potential of such models in overcoming current bottlenecks, such as weak generalization on unseen data and high computational load. This study aims to assist researchers in attaining significant insights into the state-of-the-art sequence, attention-based, hybrid, and generative models used in MI classification, thus offering a direction for future innovation.