Computer Vision Laboratory
  • Print

Image of Robail Yasrab

Robail Yasrab

Research Fellow,


Research Summary

My current research at Computer Vision Lab, University of Nottingham is focused on application of Artificial Intelligence and Machine Learning to plant phenotyping and segmentation. I am working on… read more

Selected Publications

MSc Project Ideas

1- Deep Learning in ophthalmology (detection and classification of eye disorders)

Basic knowledge of deep learning.

Artificial intelligence (AI) based on deep learning (DL) methods has sparked incredible global interest in recent years. DL has been extensively used for many problems (speech/image recognition, natural language processing, etc.) Recently, AI and DL started to impact on healthcare. In ophthalmology (the diagnosis and treatment of eye disorders), DL has been applied primarily to medical imaging analysis. There are numerous ophthalmic diseases in which deep learning methods have been used (e.g., diabetic retinopathy (DR), glaucoma, age-related macular degeneration (AMD) and retinopathy of prematurity (ROP)).

We are looking to design and develop a single DL ophthalmic system with multiple eye disease diagnosis and risk analysis. While you are encouraged to experiment and build your approach, the project will likely need to be broken into two key stages:

  • Review and preparation of custom dataset from publically available ophthalmic datasets.
  • Detection and classification of the diabetic retinopathy and retinopathy of prematurity, the glaucoma-like disc, macular oedema and age-related macular degeneration.

2- Deep Learning in Plant Phenotyping (forecasting plant growth)

Basic knowledge of deep learning.

Deep learning-based systems for plant phenotyping is more efficient for measuring different plant traits for diverse genetic discoveries compared to the traditional image-based phenotyping approaches. Plant biologists have recently demanded more reliable and accurate image-based phenotyping systems for assessing various features of plants and crops. The core of these image-based phenotyping systems is structural classification and features segmentation. Deep learning-based networks, however, have shown outstanding results in extracting very complicated features and structures of above and below ground plants.

The proposed idea is to design and develop a machine learning-based system that makes use of plant image-based datasets grown under controlled environmental conditions to forecast the pattern of future plant growth. This project will include:

  • Dataset generation and annotation an using already available Root Phenotyping tool.
  • Training Convolution network (CNN)/ Generative Adversarial Network (GAN) to map the plant growth patterns
  • Forecasting plant growth

3- Fighting Deepfakes

Basic knowledge of deep learning/machine learning.

The fake videos produced through well-known AI systems such as DeepFakes are getting more and more realistic. Currently, it becomes hard for a human to differentiate between fake and real video clips. Such videos could be harmful or misleading for someone's personal life, business or brand. Therefore, it is crucial to detect such synthetically generated fake data. The key idea of this project to analyze such information and categorize it into fake or not fake.

  • This project will involve understanding and analysis of synthetic data
  • Design and develop algorithms to detect fake video
  • Develop an automatic approach using deep learning methods like CNNs, GANs, or Recurrent Neural Network (RNNs)

4- Understanding First-person vision

Basic knowledge of deep learning/machine learning.

Understating first-person (egocentric) vision getting interest of researchers. It is helping to understand a unique viewpoint of someone's attention, interaction and handling of objects. The idea is to make use of EPIC-KITCHENS, a large-scale egocentric video dataset ( to train machine-learning systems to classify action in video and segment these actions for "Action-Anticipation" and Action-Recognition. Further details about the dataset or challenge could be found at

Please get in touch if you have any questions about the specifics of the projects.

Current Research

My current research at Computer Vision Lab, University of Nottingham is focused on application of Artificial Intelligence and Machine Learning to plant phenotyping and segmentation. I am working on LeMuR Project; a Biotechnology and Biological Sciences Research Council funded Project. The key idea is to carry out plant root phenotyping via Learned Multi-resolution Image Segmentation.

This will result in efficient analysis and segmentation of plant roots, and help biologists analyze plant traits under different natural conditions, with the ultimate aim of improving food security for the UK. The previous research in this domain at the University of Nottingham had led to the creation of manual phenotyping classification tool called RootNav 2.0.

I have been working over the past year in developing the next generation of fully automatic phenotyping tool from 2D root images, named as ROOTNAV 2.0.

Past Research

My Post-Doctorate research work at USTC China was focused on semantic segmentation of 2D images, where I specialized in road scenes using deep learning-based convolution neural networks (CNNs). The proposed designs of the CNNs were novel approaches that have shown a real-time application of CNNs could be possible for areas like road scene understanding. The resultant network architecture could be the best choice for video-based advance driving assistance system (ADAS) for offering improved performance and enhanced results. It was a low-cost solution with efficient performance in different road and weather conditions.

Computer Vision Laboratory

The University of Nottingham
Jubilee Campus
Wollaton Road
Nottingham, NG8 1BB

telephone: +44 (0) 115 8466543