Dr. Samabia Tehsin

PhD Theme/Topic: AI-Driven Underwater Imagery Analysis for Target Detection and Identification in Military Environments

Supervisor: Dr. Samabia Tehsin, Professor of Computer Science
Contact #: 0321-5010826
Email: stehseen.buic@bahria.edu.pk
Campus/School/Dept: BUIC
RAC Approved Supervisor for Research Areas: Explainable AI, Deep Learning, Machine Learning, NLP, Generative AI

Supervisory Record:
PhD Produced: 01
PhD Enrolled: 02
MS/MPhil Produced: 18
MS/MPhil Enrolled: 01

Topic Brief Description:

Underwater environments present significant challenges for visual sensing due to light absorption, scattering, turbidity, and noise. These limitations make underwater target detection—such as identifying submarines, naval mines, unmanned underwater vehicles (UUVs), or diver threats—extremely difficult using conventional imaging.

This research aims to develop advanced AI-based underwater imagery enhancement and target identification pipelines that can operate reliably in complex military maritime conditions. The project will explore deep learning methods for image restoration, spectral correction, and noise suppression, followed by robust detection and classification modules tailored to underwater military targets.

The outcomes of this research will support defense setups by enabling improved situational awareness, threat detection, and maritime security, contributing to the development of autonomous underwater surveillance systems for Pakistan Navy and related agencies.

Research Objectives/Deliverables: 

• To develop a deep-learning-based underwater image enhancement framework that corrects color distortion, improves visibility, and suppresses scattering effects in military maritime conditions.
• To design a specialized underwater target detection and classification model for the identification of submarines, naval mines, UUVs, and diver threats using multi-modal imagery.
• To evaluate target identification accuracy under varying turbidity, illumination, and depth levels, simulating real military operational environments.

Research Questions:

• How can AI-based enhancement techniques reduce underwater image degradation (e.g., scattering, low contrast, color attenuation) to improve visibility for military surveillance?
• What deep-learning architectures are best suited for reliable underwater target identification, considering environmental distortions and limited labeled data?
• What performance metrics and evaluation scenarios best reflect environments for underwater defense imaging systems? 

Candidate’s Eligibility Profile: 

• The applicant must have an MS/MPhil degree in Computer Science, AI, or related fields with a CGPA > 3.0.
• A solid background in machine learning, deep learning, image processing, and Computer Vision is required.
• Experience with Python is highly recommended.
• Prior exposure to computer vision and underwater imagery is a plus but not mandatory.