Research

The doctorate

Speech, denoised — the PhD in one pictureSynthesized log-mel spectrogram · thermal palette · Mar 2025— · Current

Edinburgh Napier University

PhD, Computing

Audio-visual speech enhancement within the COG-MHEAR programme.

Fully funded PhD studentship (COG-MHEAR programme). · Mar 2025—

CyberASAP Year 10

PrivAI: Privacy-preserving AI Analytics Engine for Enabling Secure Recommendations in Regulated Sectors

Edinburgh Napier University · Jun 2026—Feb 2027

How the work connects

Point at a topic to see what it connects to.Tap a topic to see what it connects to.

• Audio-visual speech enhancement

  Recovering intelligible speech from noise using what the face shows the camera as well as what the microphone hears.

  Connects toComputer visionMachine learningPrivacy-preserving AI

• Computer vision

  Image and video understanding — from lip movement to thermal frames to clinical scans.

  Connects toAudio-visual speech enhancementMedical AIArabic NLP

• Medical AI

  Deep learning for diagnosis and screening — brain, cervical, kidney, malaria.

  Connects toComputer visionMachine learning

• Machine learning

  Models, transfer learning, model compression — the shared toolbox under every other topic.

  Connects toAudio-visual speech enhancementMedical AIArabic NLPPrivacy-preserving AI

• Arabic NLP

  Sentiment analysis, handwritten recognition and sign-language translation for Arabic.

  Connects toMachine learningComputer visionIntelligent software systems

• Privacy-preserving AI

  Federated learning and homomorphic encryption so models learn without data leaving its owner.

  Connects toAudio-visual speech enhancementMachine learningIntelligent software systems

• Intelligent software systems

  Research that ships — production web and mobile systems that carry models to real users.

  Connects toPrivacy-preserving AIArabic NLP