Dr. Syed Muhammad Usman

PhD Theme/Topic: Automated Early Disease Diagnosis and Clinical Decision Support Using Biomedical Signals/ Radiological Imaging

Supervisor: Dr. Syed Muhammad Usman, Associate Professor
Contact #: 03320615294
Email: smusman.h11@bahria.edu.pk
Campus/School/Dept: BUH11/BSEAS/CS
RAC Approved Supervisor for Research Areas: Artificial Intelligence, Biomedical Signal Processing, Medical Imaging

Supervisory Record:
PhD Produced: Nil
PhD Enrolled: 02
MS/MPhil Produced: 08
MS/MPhil Enrolled: 05

Co-Supervisor: Prof. Dr. Shehzad Khalid, Senior Professor
Contact #: 03005102999
Email: shehzad@bahria.edu.pk
Campus/School/Dept: BUH11/BSEAS/CE
RAC Approved Supervisor for Research Areas: Artificial Intelligence, Biomedical Signal Processing, Medical Imaging

Supervisory Record:
PhD Produced: 06
PhD Enrolled: 03
MS/MPhil Produced: 20+
MS/MPhil Enrolled: Nil

Topic Brief Description:
This research focuses on the development of automated methods for early disease diagnosis and clinical decision support using biomedical signals and/or radiological imaging. It aims to explore advanced computational approaches for analyzing physiological signals such as EEG and ECG, and/or imaging modalities including X-ray and CT scans, to enable timely, accurate, and scalable detection of neurological, cardiovascular, and other critical conditions.

The work encourages the use of signal processing, pattern recognition, and data-driven modeling techniques to extract meaningful insights from complex healthcare data. Single modality (e.g., EEG, ECG, or radiological imaging) may be focused or explore multimodal approaches that integrate multiple data sources to enhance diagnostic performance and robustness.

Research Objectives/Deliverables:

• To develop synthetic data generation techniques for biomedical signals and/or radiological imaging to address challenges such as class imbalance, data scarcity, and variability, thereby improving the robustness and generalization of diagnostic models.
• To design and implement automated methods for accurate disease detection and/or clinical report generation using either biomedical signals (e.g., EEG, ECG) or radiological imaging (e.g., X-ray, CT), with the flexibility to focus on a single modality or multimodal integration.
• To incorporate explainability and interpretability mechanisms that enable clinically meaningful insights, ensuring that model predictions and generated outputs are transparent, trustworthy, and aligned with medical decision-making processes.

Research Questions:

• How can synthetic data generation techniques be developed and optimized for biomedical signals and/or radiological imaging to effectively mitigate class imbalance and data scarcity while preserving clinical relevance?
• How can automated methods be designed to achieve accurate disease detection and/or generate clinically meaningful reports using either biomedical signals (e.g., EEG, ECG) or radiological imaging (e.g., X-ray, CT)?
• How can explainability and interpretability be incorporated into automated systems to ensure transparent, reliable, and clinically useful insights for healthcare decision-making?

Candidate’s Eligibility Profile:

• Applicants must hold an MS/MPhil (or equivalent) degree in a relevant field with a minimum CGPA of 3.5 (or equivalent). A strong foundation in mathematics and related areas is required.
• Proficiency in Python programming is required, including experience with libraries and frameworks such as NumPy, Pandas, Scikit-learn, PyTorch/ Keras or TensorFlow. Experience in machine learning, signal processing, or medical image analysis will be considered an added advantage. Candidates should be able to work effectively in a collaborative and international research environment.
• Applicants must demonstrate strong written and spoken English communication skills. The ability to work independently, along with a commitment to teamwork and research excellence, is highly valued.