Yongmin Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Yongmin Li has a total of 210 co-authors. Here are the top co-authors by number of shared publications:
Xiaohui Liu
23 shared publications- Xiaohui Liu
Brunel University London
10 shared publications Alina Miron
9 shared publications- Alina Miron
Department of Computer Science, Brunel University, Uxbridge, UK
8 shared publications - Bashir Isa Dodo
Brunel University London
8 shared publications Djibril Kaba
8 shared publications
Latest Publications
- Open Access
- 15 June 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 331
Automated Radiology Report Generation: A Comprehensive Review
- 15 June 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 331
Open AccessAutomated Radiology Report Generation: A Comprehensive Review
- Open Access
- 6 February 2026
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 42 (18) , 1-19
Extracting Meaningful Insights from User Research Videos
- 6 February 2026
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 42 (18) , 1-19
Open AccessExtracting Meaningful Insights from User Research Videos
- 11 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-14
A Mutually Enhancement Network for Superpixel Segmentation and Classification of Hyperspectral Image
Most existing hyperspectral image (HSI) classification methods primarily focus on capturing subtle spectral variations by leveraging local spectral–spatial cues derived from patch-level representations. However, limited attention has been given to exploring the global spatial contextual correlations among pixels of HSI. In this study, we propose the superpixel segmentation and classification mutual enhancement network ( $\mathbf {S}^{2}$ CMEN), a novel framework that integrates global spatial correlations with spectral information through the mutual enhancement of superpixel segmentation and classification. Specifically, a global spatial adaptive module (GSAM) is designed to obtain the direct correlation of the global classes in HSI. It consists of an adaptive spectral-superpixel network (ASSN) and a graph convolutional network (GCN), forming a synergistic architecture that effectively captures global spatial relationships by adaptively deriving superpixel results from HSIs. Notably, GSAM offers a transferable global spatial representation for HSI tasks, enabling integration with other spectral feature extraction models. Furthermore, we develop a spatial–spectral fusion module (SSFM) to obtain comprehensive spectral features and fuse them with the extracted global spatial features. Finally, under the constraint of a unit loss, the mutual enhancement strategy (MES) can make the superpixel segmentation loss and the classification loss mutually enhance each other for better performance. We conducted extensive experiments on three public datasets. The proposed $\mathbf {S}^{2}$ CMEN achieves overall classification accuracies of 97.38%, 92.33%, and 91.38% on Indian Pines (IPs), Pavia University (PU), and Houston, respectively, consistently surpassing existing state-of-the-art methods.
- 11 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-14
A Mutually Enhancement Network for Superpixel Segmentation and Classification of Hyperspectral Image
Most existing hyperspectral image (HSI) classification methods primarily focus on capturing subtle spectral variations by leveraging local spectral–spatial cues derived from patch-level representations. However, limited attention has been given to exploring the global spatial contextual correlations among pixels of HSI. In this study, we propose the superpixel segmentation and classification mutual enhancement network ( $\mathbf {S}^{2}$ CMEN), a novel framework that integrates global spatial correlations with spectral information through the mutual enhancement of superpixel segmentation and classification. Specifically, a global spatial adaptive module (GSAM) is designed to obtain the direct correlation of the global classes in HSI. It consists of an adaptive spectral-superpixel network (ASSN) and a graph convolutional network (GCN), forming a synergistic architecture that effectively captures global spatial relationships by adaptively deriving superpixel results from HSIs. Notably, GSAM offers a transferable global spatial representation for HSI tasks, enabling integration with other spectral feature extraction models. Furthermore, we develop a spatial–spectral fusion module (SSFM) to obtain comprehensive spectral features and fuse them with the extracted global spatial features. Finally, under the constraint of a unit loss, the mutual enhancement strategy (MES) can make the superpixel segmentation loss and the classification loss mutually enhance each other for better performance. We conducted extensive experiments on three public datasets. The proposed $\mathbf {S}^{2}$ CMEN achieves overall classification accuracies of 97.38%, 92.33%, and 91.38% on Indian Pines (IPs), Pavia University (PU), and Houston, respectively, consistently surpassing existing state-of-the-art methods.
- Open Access
MMIS-Net for Retinal Fluid Segmentation and Detection
Purpose: Deep learning methods have shown promising results in the segmentation, and detection of diseases in medical images. However, most methods are trained and tested on data from a single source, modality, organ, or disease type, overlooking the combined potential of other available annotated data. Numerous small annotated medical image datasets from various modalities, organs, and diseases are publicly available. In this work, we aim to leverage the synergistic potential of these datasets to improve performance on unseen data. Approach: To this end, we propose a novel algorithm called MMIS-Net (MultiModal Medical Image Segmentation Network), which features Similarity Fusion blocks that utilize supervision and pixel-wise similarity knowledge selection for feature map fusion. Additionally, to address inconsistent class definitions and label contradictions, we created a one-hot label space to handle classes absent in one dataset but annotated in another. MMIS-Net was trained on 10 datasets encompassing 19 organs across 2 modalities to build a single model. Results: The algorithm was evaluated on the RETOUCH grand challenge hidden test set, outperforming large foundation models for medical image segmentation and other state-of-the-art algorithms. We achieved the best mean Dice score of 0.83 and an absolute volume difference of 0.035 for the fluids segmentation task, as well as a perfect Area Under the Curve of 1 for the fluid detection task. Conclusion: The quantitative results highlight the effectiveness of our proposed model due to the incorporation of Similarity Fusion blocks into the network's backbone for supervision and similarity knowledge selection, and the use of a one-hot label space to address label class inconsistencies and contradictions.
Open AccessMMIS-Net for Retinal Fluid Segmentation and Detection
Purpose: Deep learning methods have shown promising results in the segmentation, and detection of diseases in medical images. However, most methods are trained and tested on data from a single source, modality, organ, or disease type, overlooking the combined potential of other available annotated data. Numerous small annotated medical image datasets from various modalities, organs, and diseases are publicly available. In this work, we aim to leverage the synergistic potential of these datasets to improve performance on unseen data. Approach: To this end, we propose a novel algorithm called MMIS-Net (MultiModal Medical Image Segmentation Network), which features Similarity Fusion blocks that utilize supervision and pixel-wise similarity knowledge selection for feature map fusion. Additionally, to address inconsistent class definitions and label contradictions, we created a one-hot label space to handle classes absent in one dataset but annotated in another. MMIS-Net was trained on 10 datasets encompassing 19 organs across 2 modalities to build a single model. Results: The algorithm was evaluated on the RETOUCH grand challenge hidden test set, outperforming large foundation models for medical image segmentation and other state-of-the-art algorithms. We achieved the best mean Dice score of 0.83 and an absolute volume difference of 0.035 for the fluids segmentation task, as well as a perfect Area Under the Curve of 1 for the fluid detection task. Conclusion: The quantitative results highlight the effectiveness of our proposed model due to the incorporation of Similarity Fusion blocks into the network's backbone for supervision and similarity knowledge selection, and the use of a one-hot label space to address label class inconsistencies and contradictions.
- Open Access
- 28 November 2024
- Published by MDPI AG in Future Internet
- Vol. 16 (12) , 444
Detailed Image Captioning and Hashtag Generation
This article presents CapFlow, an integrated approach to detailed image captioning and hashtag generation. Based on a thorough performance evaluation, the image captioning model utilizes a fine-tuned vision-language model with Low-Rank Adaptation (LoRA), while the hashtag generation employs the keyword extraction method. We evaluated the state-of-the-art image captioning models using both traditional metrics (BLEU, METEOR, ROUGE-L, and CIDEr) and the specialized CAPTURE metric for detailed captions. The hashtag generation models were assessed using precision, recall, and F1-score. The proposed method demonstrates competitive results against larger models while maintaining efficiency suitable for real-time applications. The image captioning model outperforms the base Florence-2 model and favorably compares with larger models. The KeyBERT implementation for hashtag generation surpasses other keyword extraction methods in both accuracy and speed. This work contributes to the field of AI-assisted content analysis and generation, offering insights into the practical implementation of advanced vision-language models for detailed image understanding and relevant tag generation.
- 28 November 2024
- Published by MDPI AG in Future Internet
- Vol. 16 (12) , 444
Open AccessDetailed Image Captioning and Hashtag Generation
This article presents CapFlow, an integrated approach to detailed image captioning and hashtag generation. Based on a thorough performance evaluation, the image captioning model utilizes a fine-tuned vision-language model with Low-Rank Adaptation (LoRA), while the hashtag generation employs the keyword extraction method. We evaluated the state-of-the-art image captioning models using both traditional metrics (BLEU, METEOR, ROUGE-L, and CIDEr) and the specialized CAPTURE metric for detailed captions. The hashtag generation models were assessed using precision, recall, and F1-score. The proposed method demonstrates competitive results against larger models while maintaining efficiency suitable for real-time applications. The image captioning model outperforms the base Florence-2 model and favorably compares with larger models. The KeyBERT implementation for hashtag generation surpasses other keyword extraction methods in both accuracy and speed. This work contributes to the field of AI-assisted content analysis and generation, offering insights into the practical implementation of advanced vision-language models for detailed image understanding and relevant tag generation.
- Open Access
- 30 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (3) , 1257-1272
Multi-Center Fetal Brain Tissue Annotation (FeTA) Challenge 2022 Results
Segmentation is a critical step in analyzing the developing human fetal brain. There have been vast improvements in automatic segmentation methods in the past several years, and the Fetal Brain Tissue Annotation (FeTA) Challenge 2021 helped to establish an excellent standard of fetal brain segmentation. However, FeTA 2021 was a single center study, limiting real-world clinical applicability and acceptance. The multi-center FeTA Challenge 2022 focused on advancing the generalizability of fetal brain segmentation algorithms for magnetic resonance imaging (MRI). In FeTA 2022, the training dataset contained images and corresponding manually annotated multi-class labels from two imaging centers, and the testing data contained images from these two centers as well as two additional unseen centers. The multi-center data included different MR scanners, imaging parameters, and fetal brain super-resolution algorithms applied. 16 teams participated and 17 algorithms were evaluated. Here, the challenge results are presented, focusing on the generalizability of the submissions. Both in- and out-of-domain, the white matter and ventricles were segmented with the highest accuracy (Top Dice scores: 0.89, 0.87 respectively), while the most challenging structure remains the grey matter (Top Dice score: 0.75) due to anatomical complexity. The top 5 average Dices scores ranged from 0.81-0.82, the top 5 average 95 th percentile Hausdorff distance values ranged from 2.3-2.5mm, and the top 5 volumetric similarity scores ranged from 0.90-0.92. The FeTA Challenge 2022 was able to successfully evaluate and advance generalizability of multi-class fetal brain tissue segmentation algorithms for MRI and it continues to benchmark new algorithms.
- 30 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (3) , 1257-1272
Open AccessMulti-Center Fetal Brain Tissue Annotation (FeTA) Challenge 2022 Results
Segmentation is a critical step in analyzing the developing human fetal brain. There have been vast improvements in automatic segmentation methods in the past several years, and the Fetal Brain Tissue Annotation (FeTA) Challenge 2021 helped to establish an excellent standard of fetal brain segmentation. However, FeTA 2021 was a single center study, limiting real-world clinical applicability and acceptance. The multi-center FeTA Challenge 2022 focused on advancing the generalizability of fetal brain segmentation algorithms for magnetic resonance imaging (MRI). In FeTA 2022, the training dataset contained images and corresponding manually annotated multi-class labels from two imaging centers, and the testing data contained images from these two centers as well as two additional unseen centers. The multi-center data included different MR scanners, imaging parameters, and fetal brain super-resolution algorithms applied. 16 teams participated and 17 algorithms were evaluated. Here, the challenge results are presented, focusing on the generalizability of the submissions. Both in- and out-of-domain, the white matter and ventricles were segmented with the highest accuracy (Top Dice scores: 0.89, 0.87 respectively), while the most challenging structure remains the grey matter (Top Dice score: 0.75) due to anatomical complexity. The top 5 average Dices scores ranged from 0.81-0.82, the top 5 average 95 th percentile Hausdorff distance values ranged from 2.3-2.5mm, and the top 5 volumetric similarity scores ranged from 0.90-0.92. The FeTA Challenge 2022 was able to successfully evaluate and advance generalizability of multi-class fetal brain tissue segmentation algorithms for MRI and it continues to benchmark new algorithms.
- Open Access
- 8 August 2024
- Published by MDPI AG in Future Internet
- Vol. 16 (8) , 287
Dynamic Fashion Video Synthesis from Static Imagery
Online shopping for clothing has become increasingly popular among many people. However, this trend comes with its own set of challenges. For example, it can be difficult for customers to make informed purchase decisions without trying on the clothes to see how they move and flow. We address this issue by introducing a new image-to-video generator called FashionFlow to generate fashion videos to show how clothing products move and flow on a person. By utilising a latent diffusion model and various other components, we are able to synthesise a high-fidelity video conditioned by a fashion image. The components include the use of pseudo-3D convolution, VAE, CLIP, frame interpolator and attention to generate a smooth video efficiently while preserving vital characteristics from the conditioning image. The contribution of our work is the creation of a model that can synthesise videos from images. We show how we use a pre-trained VAE decoder to process the latent space and generate a video. We demonstrate the effectiveness of our local and global conditioners, which help preserve the maximum amount of detail from the conditioning image. Our model is unique because it produces spontaneous and believable motion using only one image, while other diffusion models are either text-to-video or image-to-video using pre-recorded pose sequences. Overall, our research demonstrates a successful synthesis of fashion videos featuring models posing from various angles, showcasing the movement of the garment. Our findings hold great promise for improving and enhancing the online fashion industry’s shopping experience.
- 8 August 2024
- Published by MDPI AG in Future Internet
- Vol. 16 (8) , 287
Open AccessDynamic Fashion Video Synthesis from Static Imagery
Online shopping for clothing has become increasingly popular among many people. However, this trend comes with its own set of challenges. For example, it can be difficult for customers to make informed purchase decisions without trying on the clothes to see how they move and flow. We address this issue by introducing a new image-to-video generator called FashionFlow to generate fashion videos to show how clothing products move and flow on a person. By utilising a latent diffusion model and various other components, we are able to synthesise a high-fidelity video conditioned by a fashion image. The components include the use of pseudo-3D convolution, VAE, CLIP, frame interpolator and attention to generate a smooth video efficiently while preserving vital characteristics from the conditioning image. The contribution of our work is the creation of a model that can synthesise videos from images. We show how we use a pre-trained VAE decoder to process the latent space and generate a video. We demonstrate the effectiveness of our local and global conditioners, which help preserve the maximum amount of detail from the conditioning image. Our model is unique because it produces spontaneous and believable motion using only one image, while other diffusion models are either text-to-video or image-to-video using pre-recorded pose sequences. Overall, our research demonstrates a successful synthesis of fashion videos featuring models posing from various angles, showcasing the movement of the garment. Our findings hold great promise for improving and enhancing the online fashion industry’s shopping experience.
- Open Access
Transforming Digital Marketing with Generative AI
The current marketing landscape faces challenges in content creation and innovation, relying heavily on manually created content and traditional channels like social media and search engines. While effective, these methods often lack the creativity and uniqueness needed to stand out in a competitive market. To address this, we introduce MARK-GEN, a conceptual framework that utilises generative artificial intelligence (AI) models to transform marketing content creation. MARK-GEN provides a comprehensive, structured approach for businesses to employ generative AI in producing marketing materials, representing a new method in digital marketing strategies. We present two case studies within the fashion industry, demonstrating how MARK-GEN can generate compelling marketing content using generative AI technologies. This proposition paper builds on our previous technical developments in virtual try-on models, including image-based, multi-pose, and image-to-video techniques, and is intended for a broad audience, particularly those in business management.
Open AccessTransforming Digital Marketing with Generative AI
The current marketing landscape faces challenges in content creation and innovation, relying heavily on manually created content and traditional channels like social media and search engines. While effective, these methods often lack the creativity and uniqueness needed to stand out in a competitive market. To address this, we introduce MARK-GEN, a conceptual framework that utilises generative artificial intelligence (AI) models to transform marketing content creation. MARK-GEN provides a comprehensive, structured approach for businesses to employ generative AI in producing marketing materials, representing a new method in digital marketing strategies. We present two case studies within the fashion industry, demonstrating how MARK-GEN can generate compelling marketing content using generative AI technologies. This proposition paper builds on our previous technical developments in virtual try-on models, including image-based, multi-pose, and image-to-video techniques, and is intended for a broad audience, particularly those in business management.
- Open Access
Segmenting Medical Images: From UNet to Res-UNet and nnUNet
This study provides a comparative analysis of deep learning models including UNet, Res-UNet, Attention Res-UNet, and nnUNet, and evaluates their performance in brain tumour, polyp, and multi-class heart segmentation tasks. The analysis focuses on precision, accuracy, recall, Dice Similarity Coefficient (DSC), and Intersection over Union (IoU) to assess their clinical applicability. In brain tumour segmentation, Res-UNet and nnUNet significantly outperformed UNet, with Res-UNet leading in DSC and IoU scores, indicating superior accuracy in tumour delineation. Meanwhile, nnUNet excelled in recall and accuracy, which are crucial for reliable tumour detection in clinical diagnosis and planning. In polyp detection, nnUNet was the most effective, achieving the highest metrics across all categories and proving itself as a reliable diagnostic tool in endoscopy. In the complex task of heart segmentation, Res-UNet and Attention Res-UNet were outstanding in delineating the left ventricle, with Res-UNet also leading in right ventricle segmentation. nnUNet was unmatched in myocardium segmentation, achieving top scores in precision, recall, DSC, and IoU. The conclusion notes that although Res-UNet occasionally outperforms nnUNet in specific metrics, the differences are quite small. Moreover, nnUNet consistently shows superior overall performance across the experiments. Particularly noted for its high recall and accuracy, which are crucial in clinical settings to minimize misdiagnosis and ensure timely treatment, nnUNet's robust performance in crucial metrics across all tested categories establishes it as the most effective model for these varied and complex segmentation tasks.
Open AccessSegmenting Medical Images: From UNet to Res-UNet and nnUNet
This study provides a comparative analysis of deep learning models including UNet, Res-UNet, Attention Res-UNet, and nnUNet, and evaluates their performance in brain tumour, polyp, and multi-class heart segmentation tasks. The analysis focuses on precision, accuracy, recall, Dice Similarity Coefficient (DSC), and Intersection over Union (IoU) to assess their clinical applicability. In brain tumour segmentation, Res-UNet and nnUNet significantly outperformed UNet, with Res-UNet leading in DSC and IoU scores, indicating superior accuracy in tumour delineation. Meanwhile, nnUNet excelled in recall and accuracy, which are crucial for reliable tumour detection in clinical diagnosis and planning. In polyp detection, nnUNet was the most effective, achieving the highest metrics across all categories and proving itself as a reliable diagnostic tool in endoscopy. In the complex task of heart segmentation, Res-UNet and Attention Res-UNet were outstanding in delineating the left ventricle, with Res-UNet also leading in right ventricle segmentation. nnUNet was unmatched in myocardium segmentation, achieving top scores in precision, recall, DSC, and IoU. The conclusion notes that although Res-UNet occasionally outperforms nnUNet in specific metrics, the differences are quite small. Moreover, nnUNet consistently shows superior overall performance across the experiments. Particularly noted for its high recall and accuracy, which are crucial in clinical settings to minimize misdiagnosis and ensure timely treatment, nnUNet's robust performance in crucial metrics across all tested categories establishes it as the most effective model for these varied and complex segmentation tasks.
- Open Access
- 26 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 483-489
Segmenting Medical Images: From UNet to Res-UNet and nnUNet
This study provides a comparative analysis of deep learning models—UNet, Res-UNet, Attention Res-UNet, and nnUNet—evaluating their performance in brain tumour, polyp, and multi-class heart segmentation tasks. The analysis focuses on precision, accuracy, recall, Dice Similarity Coefficient (DSC), and Intersection over Union (IoU) to assess their clinical applicability. In brain tumour segmentation, Res-UNet and nnUNet significantly outperformed UNet, with Res-UNet leading in DSC and IoU scores, indicating superior accuracy in tumour delineation. Meanwhile, nnUNet excelled in recall and accuracy, which are crucial for reliable tumour detection in clinical diagnosis and planning. In polyp detection, nnUNet was the most effective, achieving the highest metrics across all categories and proving itself as a reliable diagnostic tool in endoscopy. In the complex task of heart segmentation, Res-UNet and Attention Res-UNet were outstanding in delineating the left ventricle, with Res-UNet also leading in right ventricle segmentation. nnUNet was unmatched in myocardium segmentation, achieving top scores in precision, recall, DSC, and IoU. The conclusion notes that although ResUNet occasionally outperforms nnUNet in specific metrics, the differences are quite small. Moreover, nnUNet consistently shows superior overall performance across the experiments. Particularly noted for its high recall and accuracy, which are crucial in clinical settings to minimize misdiagnosis and ensure timely treatment, nnUNet’s robust performance in crucial metrics across all tested categories establishes it as the most effective model for these varied and complex segmentation tasks.
- 26 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 483-489
Open AccessSegmenting Medical Images: From UNet to Res-UNet and nnUNet
This study provides a comparative analysis of deep learning models—UNet, Res-UNet, Attention Res-UNet, and nnUNet—evaluating their performance in brain tumour, polyp, and multi-class heart segmentation tasks. The analysis focuses on precision, accuracy, recall, Dice Similarity Coefficient (DSC), and Intersection over Union (IoU) to assess their clinical applicability. In brain tumour segmentation, Res-UNet and nnUNet significantly outperformed UNet, with Res-UNet leading in DSC and IoU scores, indicating superior accuracy in tumour delineation. Meanwhile, nnUNet excelled in recall and accuracy, which are crucial for reliable tumour detection in clinical diagnosis and planning. In polyp detection, nnUNet was the most effective, achieving the highest metrics across all categories and proving itself as a reliable diagnostic tool in endoscopy. In the complex task of heart segmentation, Res-UNet and Attention Res-UNet were outstanding in delineating the left ventricle, with Res-UNet also leading in right ventricle segmentation. nnUNet was unmatched in myocardium segmentation, achieving top scores in precision, recall, DSC, and IoU. The conclusion notes that although ResUNet occasionally outperforms nnUNet in specific metrics, the differences are quite small. Moreover, nnUNet consistently shows superior overall performance across the experiments. Particularly noted for its high recall and accuracy, which are crucial in clinical settings to minimize misdiagnosis and ensure timely treatment, nnUNet’s robust performance in crucial metrics across all tested categories establishes it as the most effective model for these varied and complex segmentation tasks.