Research
Models and research contributions
Each contribution is shown with the status recorded on the CV. Submitted, under-review, with-editor, and accepted work is not treated as published.
Medical AI
Deep learning systems for cancer detection and clinical decision support across dermatology, neuroimaging, haematology, thyroid, lung, liver, and breast imaging.
Computer Vision
Classification, segmentation, and object detection on noisy, low-contrast, and multimodal medical images.
Vision Transformers
Residual-integrated, channel-boosted, and hybrid CNN–transformer architectures, including ARiViT, ReLViT, CBViT, and related models.
Bioinformatics and Proteomics
Protein-sequence analysis, multi-tissue protein expression classification, and computational approaches to leukaemia-related prediction.
Generative AI
GANs and related generative methods for medical image quality, augmentation, and channel-boosted transformer pipelines.
Explainable AI
Interpretability for medical imaging models, including attention maps and XAI methods for breast ultrasound.
CCNN-SCD
Dermatological diagnosis of skin cancer from lesion images, where visual inspection is labour-intensive and lesion appearance is complex.
PublishedARiViT
Classification of noisy brain medical images, where overlapping intensities and image noise hinder reliable tumour analysis.
PublishedReLViT + YOLO AML
Classification and detection of acute myeloid leukemia on bone marrow images with complex structure and overlapping intensity.
PublishedISIC systematic review
Fragmented evidence on skin cancer classification and segmentation using the ISIC datasets.
PublishedXAI for breast ultrasound
Limited transparency of deep learning models used for breast cancer ultrasound interpretation.
AcceptedMAD-YOLO-CBAM-MSF
Object detection that benefits from attention and multi-scale feature representation.
SubmittedLungCT-CBT
Classification and segmentation of lung computed tomography images.
Under reviewSEL-MTP-DPC
Multi-tissue protein expression classification from sequential compositional features.
Under reviewMad-Thy-ReLViT
Classification of thyroid cancer from medical images.
With editorCBViT
Multi-class classification of liver cancer from MRI.
Applied systems and professional builds are listed separately on Projects.