Shareefunnisa Syed
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Shareefunnisa Syed has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- K. Meena
Department of Computer Science and Engineering, GITAM School of Technology, GITAM University, Bengaluru Campus, India
4 shared publications - B. J. D. Kalyani
Department of Computer Science and Engineering, Institute of Aeronautical Engineering, Hyderabad, India
2 shared publications - B. S. Deepapriya
Department of Computer Science and Engineering, Erode Sengunthar Engineering College, Erode, India
2 shared publications - Kamaleshwar Thangavilou
Department of Computer Science and Engineering, Vel Tech Rangarajan Dr Sagunthala R&D Institute of Science and Technology, Chennai, India
2 shared publications - Kolli Venkata Krishna Kishore
Department of Computer Science and Engineering, Vignan’s Foundation for Science, Technology and Research, Guntur, India
2 shared publications - Mayuri Arul Vinayakam Rajasimman
AI Division, School of Computing Science and Engineering, VIT Bhopal University, Bhopal, India
2 shared publications
Latest Publications
- Open Access
- 10 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 39364-39383
Enhancing SER Transferability With a Gated Transformer Encoder and Hybrid Loss: A Domain Adaptation Framework
Most Speech Emotion Recognition (SER) systems do not generalize well across datasets and languages due to domain shift and restricted labeled data. To address these challenges, this work introduces the Domain Adaptation Training framework to enhance the transferability of SER models across heterogeneous speech corpora. The framework leverages XLS-R pre-trained acoustic embeddings and introduces three encoder variants: Gated Transformer Encoder(GTE), Sequential Context Encoder(SCE), and Conformer Adapter Encoder(CAE), to learn domain-invariant and emotion-discriminative features. First, a hybrid multi-objective loss is formulated based on combining Cross-Entropy, Mix-up, Contrastive, Domain-Adversarial, Pseudo-Label, and Entropy losses. This unified loss optimises inter-domain alignment, intra-class compactness, and emotion separability, enabling robust performance even under severe domain mismatches or with limited supervision. Extensive experiments have been carried out on RAVDESS → EMO-DB, IEMOCAP → EMO-DB, EMO-DB → Hindi, and RAVDESS → Hindi under supervised, semi-supervised, and unsupervised settings. In all these cases, the proposed DAT framework has performed better than the baseline models, as measured by accuracy and F1-score. Overall, the results indicate that multi objective domain adaptation significantly enhances both cross-corpus and cross-lingual SER, providing a more robust and transferable solution for real-world applications.
- 10 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 39364-39383
Open AccessEnhancing SER Transferability With a Gated Transformer Encoder and Hybrid Loss: A Domain Adaptation Framework
Most Speech Emotion Recognition (SER) systems do not generalize well across datasets and languages due to domain shift and restricted labeled data. To address these challenges, this work introduces the Domain Adaptation Training framework to enhance the transferability of SER models across heterogeneous speech corpora. The framework leverages XLS-R pre-trained acoustic embeddings and introduces three encoder variants: Gated Transformer Encoder(GTE), Sequential Context Encoder(SCE), and Conformer Adapter Encoder(CAE), to learn domain-invariant and emotion-discriminative features. First, a hybrid multi-objective loss is formulated based on combining Cross-Entropy, Mix-up, Contrastive, Domain-Adversarial, Pseudo-Label, and Entropy losses. This unified loss optimises inter-domain alignment, intra-class compactness, and emotion separability, enabling robust performance even under severe domain mismatches or with limited supervision. Extensive experiments have been carried out on RAVDESS → EMO-DB, IEMOCAP → EMO-DB, EMO-DB → Hindi, and RAVDESS → Hindi under supervised, semi-supervised, and unsupervised settings. In all these cases, the proposed DAT framework has performed better than the baseline models, as measured by accuracy and F1-score. Overall, the results indicate that multi objective domain adaptation significantly enhances both cross-corpus and cross-lingual SER, providing a more robust and transferable solution for real-world applications.
- Retraction
- 5 January 2026
- Published by Springer Nature in Soft Computing
- Vol. 30 (S1) , 639
Retraction Note: Deep neural network with reduced feature for classification of breast cancer mammogram
- 5 January 2026
- Published by Springer Nature in Soft Computing
- Vol. 30 (S1) , 639
RetractionRetraction Note: Deep neural network with reduced feature for classification of breast cancer mammogram
- RetractionOpen Access
- 23 September 2024
- Published by Springer Nature in Soft Computing
- Vol. 28 (S1) , 397
Retraction Note: Smart cataract detection system with bidirectional LSTM
- 23 September 2024
- Published by Springer Nature in Soft Computing
- Vol. 28 (S1) , 397
RetractionOpen AccessRetraction Note: Smart cataract detection system with bidirectional LSTM
- Open Access
- 14 November 2023
- Published by Qeios Ltd
Review of: "An Improved Hybrid Transfer Learning-Based Deep Learning Model for Alzheimer’s Disease Detection Using CT and MRI Scans"
ORIGINALITY AND CONTRIBUTION TO THE FIELD: The article presents an innovative approach to Alzheimer's Disease detection using a hybrid transfer learning model. It is essential to highlight how this study differs from existing research, particularly i...
- 14 November 2023
- Published by Qeios Ltd
Open AccessReview of: "An Improved Hybrid Transfer Learning-Based Deep Learning Model for Alzheimer’s Disease Detection Using CT and MRI Scans"
ORIGINALITY AND CONTRIBUTION TO THE FIELD: The article presents an innovative approach to Alzheimer's Disease detection using a hybrid transfer learning model. It is essential to highlight how this study differs from existing research, particularly i...
- Retracted
- 6 February 2023
- Published by Springer Nature in Soft Computing
- Vol. 27 (11) , 7525-7533
RETRACTED ARTICLE: Smart cataract detection system with bidirectional LSTM
- 6 February 2023
- Published by Springer Nature in Soft Computing
- Vol. 27 (11) , 7525-7533
RetractedRETRACTED ARTICLE: Smart cataract detection system with bidirectional LSTM
- Retracted
- 5 October 2022
- Published by Springer Nature in Soft Computing
- Vol. 26 (24) , 14021-14028
RETRACTED ARTICLE: Deep neural network with reduced feature for classification of breast cancer mammogram
- 5 October 2022
- Published by Springer Nature in Soft Computing
- Vol. 26 (24) , 14021-14028
RetractedRETRACTED ARTICLE: Deep neural network with reduced feature for classification of breast cancer mammogram