Maria Aouani

Ms. Maria Aouani

Computer Science, Al-Farabi Kazakh National University, Almaty, Kazakhstan
orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

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

Representative Publications

View All Publications
20 May 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ZBZholdas Buribayev
Zholdas Buribayev
MAMaria Aouani
Maria Aouani
ZZZhansaya Zhangabay
Zhansaya Zhangabay
AYAinur Yerkos
Ainur Yerkos
ZAZemfira Abdirazak
Zemfira Abdirazak
MZMukhtar Zhassuzak
Mukhtar Zhassuzak
Sign languages are characterized by complex and subtle hand movements, which are challenging for computer vision systems to accurately recognize. This study suggests an innovative deep learning pipeline specifically designed for reliable gesture recognition of Kazakh Sign Language. This approach combines key point detection using the YOLO model, optical flow estimation, and a bidirectional long short-term memory (BiLSTM) network. At the initial stage, a dataset is generated using MediaPipe, which is then used to train the YOLO model in order to accurately identify key hand points. After training, the YOLO model extracts key points and bounding boxes from video recordings of gestures, creating consistent representations of movements. To improve the recognition of dynamic gestures, the optical flow is calculated in an area covering 10% of the area around key points, which allows the dynamics of movements to be captured and provides additional time characteristics. The BiLSTM network is trained on multimodal input that combines data on keypoints, bounding boxes, and optical flow, resulting in improved gesture classification accuracy. The experimental results demonstrate that the proposed approach is superior to traditional methods based solely on key points, especially in recognizing fast and complex gestures. The proposed structure promotes the development of sign language recognition technologies, especially for poorly studied languages such as Kazakh, paving the way to more inclusive and effective communication tools.

14 October 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ZBZholdas Buribayev
Zholdas Buribayev
MZMukhtar Zhassuzak
Mukhtar Zhassuzak
MAMaria Aouani
Maria Aouani
ZZZhansaya Zhangabay
Zhansaya Zhangabay
ZAZemfira Abdirazak
Zemfira Abdirazak
AYAinur Yerkos
Ainur Yerkos
The optimization capabilities of Kolmogorov–Arnold Networks (KANs) remain largely unexplored, which has limited their practical use in video anomaly recognition compared to conventional 3D-CNNs. To address this gap, we introduce a novel hybrid optimization framework that combines a Minimax Ant System (MMAS) for hyperparameter selection with a modified DARTS strategy for adaptive tuning of the 3D KAN architecture. Unlike existing approaches, our method simultaneously optimizes both learning dynamics and architectural configurations, enabling KANs to better exploit their expressive power in spatiotemporal feature extraction. Applied to a three-class video dataset, the proposed approach improved model accuracy to 87%, surpassing the performance of a standard 3D-CNN by 6%.

18 May 2026 in Frontiers in Artificial Intelligence
Verified
Journal Article
Open access
PDF file
MZMukhtar Zhassuzak
Mukhtar Zhassuzak
ZBZholdas Buribayev
Zholdas Buribayev
MAMaria Aouani
Maria Aouani
AYAinur Yerkos
Ainur Yerkos
AYAigerim Yerimbetova
Aigerim Yerimbetova
Introduction: This study presents a bidirectional communication system designed to enhance interaction between hearing-impaired and hearing individuals using gesture recognition. Methods: The proposed framework integrates multiple components, including Kazakh Sign Language (KSL) gesture detection, closed-set sentence classification, speech recognition, and speech generation. The system is trained and evaluated using a dataset consisting of images and video recordings of KSL gestures together with audio data for Kazakh speech synthesis. Results: Experimental evaluation demonstrates that the overall system achieves a sentence classification accuracy of 92% across 12 closed-set sentence classes, while individual model accuracy did not fall below 87%. Compared to other approaches, the proposed method achieved competitive results in terms of accuracy on the collected dataset; however, inference speed comparisons are module-specific and measured as latency in ms/sample. Discussion: The results confirm the feasibility of the proposed approach as a proof-of-concept for improving accessibility and communication through automated sign language understanding. The system demonstrates practical applicability in bridging the communication gap between signers and non-signers, thereby promoting greater accessibility.