A5080765299- Colorectal Cancer Screening and Detection
- Medical Image Segmentation Techniques
- Esophageal Cancer Research and Treatment
- Radiomics and Machine Learning in Medical Imaging
- Gastric Cancer Management and Outcomes
- AI in cancer detection
- Advanced Image and Video Retrieval Techniques
- Gastrointestinal Bleeding Diagnosis and Treatment
- Image Retrieval and Classification Techniques
- Advanced Vision and Imaging
- Surgical Simulation and Training
- Lung Cancer Diagnosis and Treatment
- Medical Imaging and Analysis
- Retinal Imaging and Analysis
- COVID-19 diagnosis using AI
- Esophageal and GI Pathology
- Advanced Neural Network Applications
- Retinal Diseases and Treatments
- Robotics and Sensor-Based Localization
- Augmented Reality Applications
- Advanced X-ray and CT Imaging
- Bladder and Urothelial Cancer Treatments
- Mycobacterium research and diagnosis
- Artificial Intelligence in Healthcare and Education
- Cutaneous Melanoma Detection and Management
University of Leeds
2022-2024
University of Engineering and Technology Lahore
2024
Wilfrid Laurier University
2024
University of Oxford
2019-2023
National Institute for Health Research
2021-2023
NIHR Biomedical Research Centre at The Royal Marsden and the ICR
2021-2023
Oxford BioMedica (United Kingdom)
2020-2022
Health Data Research UK
2021-2022
Open Data Institute
2020-2021
OsloMet – Oslo Metropolitan University
2021
Computer-aided detection, localisation, and segmentation methods can help improve colonoscopy procedures. Even though many have been built to tackle automatic detection of polyps, benchmarking state-of-the-art still remains an open problem. This is due the increasing number researched computer vision that be applied polyp datasets. Benchmarking novel provide a direction development automated tasks. Furthermore, it ensures produced results in community are reproducible fair comparison...
The increase of available large clinical and experimental datasets has contributed to a substantial amount important contributions in the area biomedical image analysis. Image segmentation, which is crucial for any quantitative analysis, especially attracted attention. Recent hardware advancement led success deep learning approaches. However, although models are being trained on datasets, existing methods do not use information from different epochs effectively. In this work, we leverage...
The Endoscopy Computer Vision Challenge (EndoCV) is a crowd-sourcing initiative to address eminent problems in developing reliable computer aided detection and diagnosis endoscopy systems suggest pathway for clinical translation of technologies. Whilst widely used diagnostic treatment tool hollow-organs, there are several core challenges often faced by endoscopists, mainly: 1) presence multi-class artefacts that hinder their visual interpretation, 2) difficulty identifying subtle...
Abstract Polyps in the colon are widely known cancer precursors identified by colonoscopy. Whilst most polyps benign, polyp’s number, size and surface structure linked to risk of cancer. Several methods have been developed automate polyp detection segmentation. However, main issue is that they not tested rigorously on a large multicentre purpose-built dataset, one reason being lack comprehensive public dataset. As result, may generalise different population datasets. To this extent, we...
Abstract Polyps are well-known cancer precursors identified by colonoscopy. However, variability in their size, appearance, and location makes the detection of polyps challenging. Moreover, colonoscopy surveillance removal highly operator-dependent procedures occur a complex organ topology. There exists high missed rate incomplete colonic polyps. To assist clinical reduce rates, automated methods for detecting segmenting using machine learning have been achieved past years. major drawback...
Endoscopy is a routine imaging technique used for both diagnosis and minimally invasive surgical treatment. Artifacts such as motion blur, bubbles, specular reflections, floating objects pixel saturation impede the visual interpretation automated analysis of endoscopy videos. Given widespread use in different clinical applications, robust reliable identification artifacts restoration corrupted video frames fundamental medical problem. Existing state-of-the-art methods only deal with...
Abstract We present a comprehensive analysis of the submissions to first edition Endoscopy Artefact Detection challenge (EAD). Using crowd-sourcing, this initiative is step towards understanding limitations existing state-of-the-art computer vision methods applied endoscopy and promoting development new approaches suitable for clinical translation. routine imaging technique detection, diagnosis treatment diseases in hollow-organs; esophagus, stomach, colon, uterus bladder. However nature...
Deep learning in gastrointestinal endoscopy can assist to improve clinical performance and be helpful assess lesions more accurately. To this extent, semantic segmentation methods that perform automated real-time delineation of a region-of-interest, e.g., boundary identification cancer or pre-cancerous lesions, benefit both diagnosis interventions. However, accurate endoscopic images is extremely challenging due its high operator dependence high-definition image quality. utilize settings, it...
Widely used traditional supervised deep learning methods require a large number of training samples but often fail to generalize on unseen datasets. Therefore, more general application any trained model is quite limited for medical imaging clinical practice. Using separately models each unique lesion category or patient population will sufficiently curated datasets, which not practical use in real-world set-up. Few-shot approaches can only minimize the need an enormous reliable ground truth...
Methods based on convolutional neural networks have improved the performance of biomedical image segmentation. However, most these methods cannot efficiently segment objects variable sizes and train small biased datasets, which are common for use cases. While exist that incorporate multi-scale fusion approaches to address challenges arising with sizes, they usually complex models more suitable general semantic segmentation problems. In this paper, we propose a novel architecture called...
Precise instrument segmentation aids surgeons to navigate the body more easily and increases patient safety. While accurate tracking of surgical instruments in real-time plays a crucial role minimally invasive computer-assisted surgeries, it is challenging task achieve, mainly due to: (1) complex environment, (2) model design trade-off terms both optimal accuracy speed. Deep learning gives us opportunity learn environment from large surgery scene environments placements these real world...
Gastrointestinal (GI) endoscopy has been an active field of research motivated by the large number highly lethal GI cancers. Early cancer precursors are often missed during endoscopic surveillance. The high rate such abnormalities is thus a critical bottleneck. Lack attentiveness due to tiring procedures, and requirement training few contributing factors. An automatic disease classification system can help reduce risks flagging suspicious frames lesions. consists several multi-organ...
In real-world data, long-tailed data distribution is common, making it challenging for models trained on empirical risk minimisation to learn and classify tail classes effectively. While many studies have sought improve long recognition by altering the in feature space adjusting model decision boundaries, research synergy corrective approach among various methods limited. Our study delves into three long-tail techniques: Supervised Contrastive Learning (SCL), Rare-Class Sample Generator...
Ultrasound (US) imaging is clinically invaluable due to its noninvasive and safe nature. However, interpreting US images challenging, requires significant expertise, time, often prone errors. Deep learning offers assistive solutions such as segmentation. Supervised methods rely on large, high-quality, consistently labeled datasets, which are challenging curate. Moreover, these tend underperform out-of-distribution data, limiting their clinical utility. Self-supervised (SSL) has emerged a...
Gastrointestinal (GI) pathologies are periodically screened, biopsied, and resected using surgical tools. Usually, the procedures treated or areas not specifically tracked analysed during after colonoscopies. Information regarding disease borders, development amount size of area get lost.This can lead to poor follow-up bothersome reassessment difficulties post-treatment. To improve current standard also foster more research on topic we have released "Kvasir-Instrument" dataset which consists...
Data-driven methods have shown tremendous progress in medical image analysis. In this context, deep learning-based supervised are widely popular. However, they require a large amount of training data and face issues generalisability to unseen datasets that hinder clinical translation. Endoscopic imaging is characterised by inter- intra-patient variability makes these models more challenging learn representative features for downstream tasks. Thus, despite the publicly available can be...