A5105462955- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Medical Image Segmentation Techniques
- Video Analysis and Summarization
- Advanced Data Compression Techniques
- Metaheuristic Optimization Algorithms Research
- Image Enhancement Techniques
- AI in cancer detection
- Neural Networks and Applications
- Color Science and Applications
- Cutaneous Melanoma Detection and Management
- Infrared Thermography in Medicine
- Fuzzy Logic and Control Systems
- Evolutionary Algorithms and Applications
- Optical Coherence Tomography Applications
- Advanced Image Fusion Techniques
- Digital Media Forensic Detection
- Advanced Vision and Imaging
- Image and Signal Denoising Methods
- Digital Imaging for Blood Diseases
- Retinal Imaging and Analysis
- Imbalanced Data Classification Techniques
- Thermography and Photoacoustic Techniques
- Industrial Vision Systems and Defect Detection
- Face and Expression Recognition
Loughborough University
2015-2024
Islamic Azad University, Tehran
2021
École de Technologie Supérieure
2021
Université de Montréal
2021
National Cheng Kung University
2021
Tel Aviv University
2021
Southern University of Science and Technology
2021
Otto-von-Guericke University Magdeburg
2021
Brunel University of London
2021
University of Glasgow
2021
Standardised image databases or rather the lack of them are one main weaknesses in field content based retrieval (CBIR). Authors often use their own images do not specify source datasets. Naturally this makes comparison results somewhat difficult. While a first approach towards common colour set has been taken by MPEG 7 committee database does cater for all strands research CBIR community. In particular as MPEG-7 only exist compressed form it allow an objective evaluation algorithms that...
Generalized nucleus segmentation techniques can contribute greatly to reducing the time develop and validate visual biomarkers for new digital pathology datasets. We summarize results of MoNuSeg 2018 Challenge whose objective was generalizable nuclei in pathology. The challenge an official satellite event MICCAI conference which 32 teams with more than 80 participants from geographically diverse institutes participated. Contestants were given a training set 30 images seven organs annotations...
Skin cancer is one of the major types cancers with an increasing incidence over past decades. Accurately diagnosing skin lesions to discriminate between benign and malignant crucial ensure appropriate patient treatment. While there are many computerised methods for lesion classification, convolutional neural networks (CNNs) have been shown be superior classical methods. In this work, we propose a fully automatic method classification which employs optimised deep features from number...
Automated water body detection from satellite imagery is a fundamental stage for urban hydrological studies. In recent years, various deep convolutional neural network (DCNN)-based methods have been proposed to segment remote sensing data collected by conventional RGB or multispectral such However, how effectively explore the wider spectrum bands of sensors achieve significantly better performance compared use only has left underexplored. this article, we propose novel DCNN...
Dermoscopy is one of the major imaging modalities used in diagnosis melanoma and other pigmented skin lesions. Due to difficulty subjectivity human interpretation, automated analysis dermoscopy images has become an important research area. Border detection often first step this analysis. In many cases, lesion can be roughly separated from background using a thresholding method applied blue channel. However, no single appears robust enough successfully handle wide variety encountered clinical...
Image segmentation is an important task in analysing dermoscopy images as the extraction of borders skin lesions provides cues for accurate diagnosis. One family algorithms based on idea clustering pixels with similar characteristics. Fuzzy c-means has been shown to work well segmentation, however due its iterative nature this approach excessive computational requirements. In paper, we introduce a new mean shift fuzzy algorithm that requires less time than previous techniques while providing...
This paper considers the problem of maximizing number task allocations in a distributed multirobot system under strict time constraints, where other optimization objectives need also be considered. It builds upon existing allocation algorithms, extending them with novel method for assignments. The fundamental idea is that assignment to robot has high cost if its reassignment another creates feasible slot unallocated tasks. Multiple reassignments among networked robots may required create and...
Nuclei instance segmentation plays an important role in the analysis of Hematoxylin and Eosin (H&E)-stained images. While supervised deep learning (DL)-based approaches represent state-of-the-art automatic nuclei segmentation, annotated datasets are required to train these models. There two main types tissue processing protocols, namely formalin-fixed paraffin-embedded samples (FFPE) frozen (FS). Although FFPE-derived H&E stained sections most widely used samples, staining on derived from FS...
Conventional steganography approaches embed a secret message into carrier for concealed communication but are prone to attack by recent advanced steganalysis tools. In this paper, we propose Image DisEntanglement Autoencoder Steganography (IDEAS) as novel without embedding (SWE) technique. Instead of directly the image, our approach hides it transforming synthesised and is thus fundamentally immune typical attacks. By disentangling an image two representations structure texture, exploit...
Weakly supervised semantic segmentation (WSSS) has gained significant popularity as it relies only on weak labels such image level annotations rather than the pixel required by (SSS) methods. Despite drastically reduced annotation costs, typical feature representations learned from WSSS are representative of some salient parts objects and less reliable compared to SSS due guidance during training. In this paper, we propose a novel Multi-Strategy Contrastive Learning (MuSCLe) framework obtain...
Plagiarism detection is a challenging task, aiming to identify similar items in two documents. In this paper, we present novel approach automatic plagiarism that combines BERT (bidirectional encoder representations from transformers) word embedding, attention mechanism-based long short-term memory (LSTM) networks, and an improved differential evolution (DE) algorithm for weight initialisation. used pretrain deep bidirectional all layers, while the pre-trained model can be fine-tuned with...
Foot ulcer is a common complication of diabetes mellitus and, associated with substantial morbidity and mortality, remains major risk factor for lower leg amputations. Extracting accurate morphological features from foot wounds crucial appropriate treatment. Although visual inspection by medical professional the approach diagnosis, this subjective error-prone, computer-aided approaches thus provide an interesting alternative. Deep learning-based methods, in particular convolutional neural...