Skip to main content

AI in Medicine and Surgery (AIMS group)

AI in Medicine and Surgery (AIMS group)

Dr Sharib Ali

Lecturer/Researcher in Artificial Intelligence/Computer vision
(Group leader)
PhD and Master's by research in Computer Vision

Pedro Chavarrias Solano

PhD research student

Gerardo Loza Galindo

PhD research student

Raneem Toman

PhD research student

Edward Ellis

PhD research student

Mansoor Ali Teevno

Visiting PhD research student

Current Research/PhD co-supervision

Ziang Xu - University of Oxford, UK
Xukun Zhang - Fudan University, China
Aya Hammad - University of York, York, UK
Maksim Richards - University of Oxford, Oxford, UK
Ajay Patha - NAAMII, Kathmandu, Nepal
Helena Valencia - Tec de Monterrey, Mexico

Past students

Darshita Budhadev, MRes graduate, University of Leeds (Now AI Researcher at Hartree Centre, STFC Daresbury, UK)
Soumya Gupta, DPhil graduate from University of Oxford (Now Research Engineer at Vision RT Ltd, London, UK)
Juan Carlos Ángeles Cerón, Master's graduate from Tec de Monterrey (Now Data and Applied Scientist II at Microsoft, Washington, US)
Shruti Shrestha, research intern (Now Machine Learning Research Scientist @ Georgia State University, Atlanta, US)
Nikhil K Tomar, research intern (Now working with collaborators at the Northwestern University, Chicago, US)

Open positions in the group

Project: Federated learning for tackling multimodal and class-imbalance problems in healthcare

A highly competitive EPSRC Doctoral Training Partnership Studentship offering the award of fees, together with a tax-free maintenance grant (currently £18,622 for academic session 2023/24) for 3.5 years.  Training and support will also be provided.

This opportunity is open to all applicants.  All candidates will be placed into the EPSRC Doctoral Training Partnership Studentship Competition and selection is based on academic merit.

To submit please follow the instructions at this link here: https://phd.leeds.ac.uk/project/1851-federated-learning-for-tackling-multimodal-and-class-imbalance-problems-in-healthcare 

Closing date: 19 February 2024

Project: Artificial Intelligence for Wearables data-analysis for personalised cancer treatment

A fully funded PhD for four years (through MRC DiMeN Doctoral Training Partnership) is available at the University of Leeds

Please click here for more details.

Further information on the programme and how to apply can be found on our website:

https://www.dimen.org.uk/how-to-apply

Closing date: 18 December 2023

 

News!!!

 

Research highlights

Assessing generalisability of deep learning-based in colonoscopy

To assist in clinical procedures and reduce missed rates, automated methods for detecting and segmenting polyps using machine learning have been achieved in past years. However, the major drawback in most of these methods is their ability to generalise to out-of-sample unseen datasets from different centres, populations, modalities, and acquisition systems. To test this hypothesis rigorously, we, together with expert gastroenterologists, curated a multi-centre and multi-population dataset acquired from six different colonoscopy systems and challenged the computational expert teams to develop robust automated detection and segmentation methods in a crowd-sourcing Endoscopic computer vision challenge. This work put forward rigorous generalisability tests and assesses the usability of devised deep learning methods in dynamic and actual clinical colonoscopy procedures.

Where do we stand in AI for endoscopic image analysis?

The paper reviews recent works on endoscopic image analysis with artificial intelligence (AI) and emphasises the current unmatched needs in this field. Finally, it outlines the future directions for clinically relevant complex AI solutions to improve patient outcomes. Published at npj Digit. Med. (December 2022) - Click to read more.

Real-time surgical tool detection with multi-scale positional encoding & CL

We propose an anchor-free architecture based on a transformer that utilises multi-scale features within the feature extraction layer and at the transformer-based detection architecture through positional encoding that can refine and capture context-aware and structural information of different-sized tools. Furthermore, a supervised contrastive loss is introduced to optimize representations of object embeddings, resulting in improved feed-forward network performances for classifying localized bounding boxes. Compared to the most accurate existing SOTA method, our approach has an improvement of nearly 4% on mAP50 and a reduction in the inference time by 113%. It also showed a 7% higher mAP50 than the baseline model.

Anatomical-aware Point-Voxel Network for Couinaud Segmentation in Liver CT

we design a multi-scale point-voxel fusion network to capture the anatomical structure and semantic information of the liver and vessels, respectively, while also increasing important data access through vessel structure prior. Finally, the network outputs the classification of Couinaud segments in the continuous liver space, producing a more accurate and smooth 3D Couinaud segmentation mask. Our proposed method outperforms several state-of-the-art methods, both point-based and voxel-based, as demonstrated by our experimental results on two public liver datasets.

Patch-level instance-group discrimination for colitis scoring

Inflammatory bowel disease (IBD), in particular ulcerative colitis (UC), is graded by endoscopists and this assessment is the basis for risk stratification and therapy monitoring. Presently, endoscopic characterisation is largely operator dependant leading to sometimes undesirable clinical outcomes for patients with IBD. Most existing deep learning classification methods cannot detect these fine-grained changes which make UC grading such a challenging task. In this work, we introduce a novel patch-level instance-group discrimination with pretext-invariant representation learning (PLD-PIRL) for self-supervised learning (SSL).

A semi-supervised Teacher-Student framework for surgical tool localisation

Semi-supervised learning (SSL) has recently emerged as a viable alternative showing promise in producing models retaining competitive performance to supervised methods. This paper introduces an SSL framework in the surgical tool detection paradigm, which aims to mitigate training data scarcity and data imbalance problems through a knowledge distillation approach.

Selected/latest publications

Conference (selected)

  • Xukun Zhang, Yang Liu, Sharib Ali, and others. Anatomical-aware Point-Voxel Network for Couinaud Segmentation in Liver CT. Medical Image Computing and Computer Assisted Intervention – MICCAI 2023.
  • Matthias Eisenmann, Annika Reinke, Vivienn Weru, Minu Dietlinde Tizabi, Fabian Isensee, Tim J Adler, Sharib Ali and others. Why is the winner the best? Proceedings of the  IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2023.
  • Rafael Martínez-García-Peña, Mansoor Ali Teevno, Gilberto Ochoa-Ruiz, and Sharib Ali.
    SUPRA: Superpixel Guided Loss for Improved Multi-Modal Segmentation in Endoscopy. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2023, pp. 285-294
  • Nikhil Kumar Tomar, Debesh Jha, Ulas Bagci, and Sharib Ali. TGANet: Text-guided attention for improved polyp segmentation, Medical Image Computing and Computer Assisted Intervention – MICCAI 2022. MICCAI 2022, pp. 151–160.
  • Abhishek Srivastava, Sukulpa Chanda, Debesh Jha, Umapada Pal and Sharib Ali, "GMSRF-Net: An Improved generalizability with Global Multi-Scale Residual Fusion Network for Polyp Segmentation," 2022 26th International Conference on Pattern Recognition (ICPR), Montreal, QC, Canada, 2022, pp. 4321-4327.
  • Pedro E. Chavarrias-Solano, Mansoor A. Teevno, Gilberto Ochoa-Ruiz and Sharib Ali. Knowledge Distillation with a Class-Aware Loss for Endoscopic Disease Detection. Cancer Prevention Through Early Detection. CaPTion 2022. Lecture Notes in Computer Science, vol 13581, pp. 67–76
  • Soumya Gupta, Sharib Ali, Ziang Xu, Binod Bhattarai, Ben Turney and Jens Rittscher. UNet-eVAE: Iterative Refinement Using VAE Embodied Learning for Endoscopic Image Segmentation. Machine Learning in Medical Imaging. MLMI 2022. Lecture Notes in Computer Science, vol 13583, pp. 161–170.
  • Ziang Xu, Sharib Ali, Soumya Gupta et al. (2022). Patch-Level Instance-Group Discrimination with Pretext-Invariant Learning for Colitis Scoring. In: Machine Learning in Medical Imaging. MLMI 2022. Lecture Notes in Computer Science, vol 13583. Springer, Cham. https://doi.org/10.1007/978-3-031-21014-3_11

Journals (selected)

  • Sharib Ali, Noha Ghatwary, Debesh Jha et al. Assessing generalisability of deep learning-based polyp detection and segmentation methods through a computer vision challenge. Sci Rep 14, 2032 (2024). https://doi.org/10.1038/s41598-024-52063-x (New article)
  • Gerardo Loza,  Pietro Valdastri,  Sharib Ali. Real-time surgical tool detection with multi-scale positional encoding and contrastive learning. Healthc. Technol. Lett. 00, 1–11 (2023). https://doi.org/10.1049/htl2.12060
  • Sharib Ali, Debesh Jha, Noha Ghatwary, Stefano Realdon and others. A multi-centre polyp detection and segmentation dataset for generalisability assessment. Scientific Data. 2023; 10(1):75. https://doi.org/10.1038/s41597-023-01981-y
  • Sharib Ali. Where do we stand in AI for endoscopic image analysis? Deciphering gaps and future directions. npj Digital Medicine. 2022; 5:184. https://doi.org/10.1038/s41746-022-00733-3
  • Mansoor Ali, Gilberto Ochoa-Ruiz and Sharib Ali. A semi-supervised Teacher-Student framework for surgical tool detection and localization. Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization. 2022. https://doi.org/10.1080/21681163.2022.2150688
  • Juan Carlos Angeles Ceron, Gilberto Ochoa Ruiz, Leonardo Chang and Sharib Ali. Real-time instance segmentation of surgical instruments using attention and multi-scale feature fusion. Medical Image Analysis. 2022; 81:102569. https://doi.org/10.1016/j.media.2022.102569
  • Soumya Gupta, Sharib Ali, Louise Goldsmith, Ben Turney and Jens Rittscher. Multi-class motion-based semantic segmentation for ureteroscopy and laser lithotripsy. Computerized Medical Imaging and Graphics. 2022;101:102112.

Grants & awards

Optimise project

Novel robust computer vision methods and synthetic datasets for minimally invasive surgery
£10,000 for 12 Months (co-I)

Bowel cancer prediction

£98,691 for 24 Months (co-I)

Collaborators

We collaborate with several clinical colleagues at the Leeds Teaching Hospital Trust  and other hospitals in the UK including Oxford University Hospitals, Oxford. We have also extended collaboration across Europe (France, Italy, Sweden, and Germany), Canada, Asia (China, Nepal)  and Africa (Egypt). Thanks to all colleagues for their continuous support in our research that we do. If your research resonate to something we do please reach out to us at s.s.ali[at]leeds[dot]ac[dot]uk