Research Experience 🔍

PROJECT(S)

Two-Step Hierarchical Binary Classification of Cancer Skin Lesions based on Transfer Learning and Random Forest

Proposed a novel two-step hierarchical binary classification framework to address skin lesion classification challenges using hybrid machine learning and deep-learning techniques.

BRAIN MRI TISSUE SEGMENTATION (IBSR18 CHALLENGE)

Deployed deep learning (U-Net architecture with ResNet34 backbone) to segment brain MRI into CSF, WM, and GM tissues, addressing challenges like spatial resolution and intensity variation.

LUNG CT IMAGE REGISTRATION (COPD CHALLENGE)

Leveraged ITK Elastix toolkit for lung CT scan registration, aligning inhale (fixed) to exhale (moving) phases from the COPDGene dataset, addressing key parameters for precise image alignment.

ADDITIONAL PROJECT(S)

Semantic Nuclei Segmentation for Medical Image Analysis using U-Net: Link
Diverse preprocessing methods on KNN model for breast cancer detection: Link
Diabetic Retinopathy Detection with Deep Learning
Image Super Resolution Using Autoencoders
Computer Aided Surgery & Medical Robotics - Bone Marrow Aspiration: Link
Boxwars-Ghana Information Platform: Link
Design of an ambulance stretcher cum wheelchair using local materials: Link
Figma Projects: MedicalAppUI | BuDbUI

PUBLICATION(S)

BuDb: A Curated Drug Discovery Database for Buruli Ulcer

Kwofie, S. K., Anyimadu, D. T., Aryee, S., Asare, B., Kokroko, N., Owusu, J. A., Afutu, B., Agyapong, O., Mosi, L., Kyei-Baffour, E., Enninful, K. S., Agoni, C., & Wilson, M. D. (2022). BuDb: A Curated Drug Discovery Database for Buruli Ulcer. https://doi.org/10.1142/S2737416523500011

BuDb is an online repository that provides enriched datasets and links to functional annotation databases that can help augment the search for new biotherapeutic entities for Buruli ulcer (BU), an ulcerative and chronic skin disease caused by Mycobacterium ulcerans.

Biotechnological drug development – The role of proteins, genes, in-silico, and stem cells in designing models for enhanced drug discovery

Suleiman, T. A., Francis, A. C., Ibrahim, A., Jonn-Joy, A. O., Adebiyi, E., & Anyimadu, D. T. (2023). Biotechnological Drug Development—The Role of Proteins, Genes, In-Silico, and Stem Cells in Designing Models for Enhanced Drug Discovery. Open Access Library Journal, 10(8), 1–20. https://doi.org/10.4236/OALIB.1110520

This review explores biotechnology's integration in drug discovery, emphasizing its efficiency in revealing new therapies compared to traditional methods. It examines the roles of proteins, genes, in-silico techniques, and stem cells in advanced drug model design, underscoring their impact on disease treatment and quality of life. The review also addresses biotech's contributions to gene therapies, cancer vaccines, and novel drug classes, concluding with a proposal for AI integration to advance drug discovery.

Two-step hierarchical binary classification of cancerous skin lesions using transfer learning and the random forest algorithm.

Suleiman, T. A., Anyimadu, D. T., Permana, A. D., Ngim, H. A. A., & Scotto di Freca, A. (2024). Two-step hierarchical binary classification of cancerous skin lesions using transfer learning and the random forest algorithm. Visual Computing for Industry, Biomedicine, and Art 2024 7:1, 7(1), 1–17. https://doi.org/10.1186/S42492-024-00166-7

This study proposes a two-step hierarchical binary classification approach for cancerous skin lesions, combining DenseNet121 for feature extraction and a random forest classifier to handle data imbalance, achieving a balanced multiclass accuracy of 91.07%. The hybrid model demonstrated improved performance and reduced computational time, showing potential for real-world applications in skin lesion classification.

CONFERENCE(S)

International Conference on Nephrology and Biotechnology, March 17-18, 2023, Chicago, U.S.A

Our research on “BuDb: A Curated Drug Discovery Database for Buruli Ulcer” was presented.

PYTHON COURSES & Relevant Link

Page updated

Google Sites

Report abuse