Search

Abstract: Medical imaging plays a vital role in medical diagnosis and clinical treatment. The biggest challenge in the medical field is the correct identification of disease and better treatment. Multi-modal Medical Image Fusion (MMIF) is the process of merging multiple medical images from different modalities into a single fused image. The main objective of the medical image fusion is to obtain a large amount of appropriate information (i.e., features) to improve the quality and make it more informative for increasing clinical therapy for better diagnosis and clear assessment of medical-related problems. The MMIF is generally considered with MRI (Magnetic Resonance Imaging), CT (Computed Tomography), PET (Positron Emission Tomography), SPECT (Single Photon Emission Computed Tomography), MRA (Magnetic Resonance Angiography), T1-weighted MR, T2-weighted MR, X-ray, and ultrasound imaging (Vibro-Acoustography). This review article presents a comprehensive survey of existing medical image fusion methods and has been characterized into six parts: (1) Multi-modality medical images, (2) Literature review process, (3) Image fusion rules, (4) Quality evaluation metrics for assessment of fused image, (5) Experimental results on registered datasets and (6) Conclusion. In addition, this review article provides scientific challenges faced in MMIF and future directions for better diagnosis. It is expected that this review will be useful in establishing a concrete foundation for developing more valuable fusion methods for medical diagnosis.

Kidney abnormality is one of the major concerns in modern society, and it affects millions of people around the world. To diagnose different abnormalities in human kidneys, a narrow-beam x-ray imaging procedure, computed tomography, is used, which creates cross-sectional slices of the kidneys. Several deep-learning models have been successfully applied to computer tomography images for classification and segmentation purposes. However, it has been difficult for clinicians to interpret the model’s specific decisions and, thus, creating a “black box” system. Additionally, it has been difficult to integrate complex deep-learning models for internet-of-medical-things devices due to demanding training parameters and memory-resource cost. To overcome these issues, this study proposed (1) a lightweight customized convolutional neural network to detect kidney cysts, stones, and tumors and (2) understandable AI Shapely values based on the Shapley additive explanation and predictive results based on the local interpretable model-agnostic explanations to illustrate the deep-learning model. The proposed CNN model performed better than other state-of-the-art methods and obtained an accuracy of 99.52 ± 0.84% for K = 10-fold of stratified sampling. With improved results and better interpretive power, the proposed work provides clinicians with conclusive and understandable results.

Breast cancer is one of the most fatal diseases leading to the death of several women across the world. But early diagnosis of breast cancer can help to reduce the mortality rate. So an efficient multi-task learning approach is proposed in this work for the automatic segmentation and classification of breast tumors from ultrasound images. The proposed learning approach consists of an encoder, decoder, and bridge blocks for segmentation and a dense branch for the classification of tumors. For efficient classification, multi-scale features from different levels of the network are used. Experimental results show that the proposed approach is able to enhance the accuracy and recall of segmentation by [Formula: see text], [Formula: see text], and classification by [Formula: see text], [Formula: see text], respectively than the methods available in the literature.

Magnitude-based pruning is a technique used to optimise deep learning models for edge inference. We have achieved over 75% model size reduction with a higher accuracy than the original multi-output regression model for head-pose estimation., Conference:- in TENCON 2022 - 2022 IEEE Region 10 Conference (TENCON)

The selection of the proper classifier and the implementation of the proper training strategy have the most impact on the performance of machine learning classifiers. The amount and distribution of data used for training and validation is another crucial aspect of classifier performance. The goal of this study was to identify the optimal combination of classifiers and validation strategies for achieving the highest accuracy rate while testing models with a small dataset. To that end, five primary classifiers were examined with varying proportions of training data and validation procedures. Most of the time, Random Forest and Nave Bayes classifier models outperformed competing classifiers. However, we discovered the best performance when we employed the holdout cross-validation technique using 70% of the available data as a training set and the remaining data as a test set.

It is an established fact that the genuineness of facial micro-expression is an effective means for estimating concealed emotions (Li et al. in Micro-expression recognition under low-resolution cases. SciTePress, Science and Technology Publications, Setúbal, 2019). Conventionally, analysis of these expressions has been performed using high resolution images which are ideal cases. However, in a real-world scenario, capturing expressions with high resolution images may not always be possible particularly using low-cost surveillance cameras. Faces captured using such cameras are often very tiny and of poor resolution. Due to the loss of discriminative features these images may not be of much use particularly for identifying certain minute facial details. To make these images useful, enhancing the textural information becomes essential and super-resolution algorithms can be ideal to achieve this. In this work, we utilize algorithms based on deep learning and generative adversarial network for transforming low-resolution micro-expression images into super-resolution images and examine their fitness particularly for micro-expression recognition. The proposed approach is tested on simulated dataset obtained from two popular spontaneous micro-expression datasets namely CASME II and SMIC-VIS; the experimental results demonstrate that the method achieved favourable results with the best recognition performance recorded as 61.63%. The significance of this work is: first, it thoroughly investigates reconstruction performance of several deep learning super-resolution algorithms on simulated low-quality micro-expression images; second, it provides a comprehensive analysis of the results obtained employing these reconstructed images to determine their contribution in addressing image quality issues specifically for micro-expression recognition.

Autism spectrum disorder is an umbrella term for a group of neurodevelopmental disorders that is associated with impairments to social interaction, communication, and behaviour. Typically, autism spectrum disorder is first detected with a screening tool (e.g. modified checklist for autism in toddlers). However, the interpretation of autism spectrum disorder behavioural symptoms varies across cultures: the sensitivity of modified checklist for autism in toddlers is as low as 25 per cent in Sri Lanka. A culturally sensitive screening tool called pictorial autism assessment schedule has overcome this problem. Low- and middle-income countries have a shortage of mental health specialists, which is a key barrier for obtaining an early autism spectrum disorder diagnosis. Early identification of autism spectrum disorder enables intervention before atypical patterns of behaviour and brain function become established. This article proposes a culturally sensitive autism spectrum disorder screening mobile application. The proposed application embeds an intelligent machine learning model and uses a clinically validated symptom checklist to monitor and detect autism spectrum disorder in low- and middle-income countries for the first time. Machine learning models were trained on clinical pictorial autism assessment schedule data and their predictive performance was evaluated, which demonstrated that the random forest was the optimal classifier (area under the receiver operating characteristic (0.98)) for embedding into the mobile screening tool. In addition, feature selection demonstrated that many pictorial autism assessment schedule questions are redundant and can be removed to optimise the screening process.

Automatic segmentation of multi-modal Cardiac Magnetic Resonance Imaging (CMRI) scans is challenging due to the variant intensity distribution and unclear boundaries between the neighbouring tissues and other organs. The deep convolutional neural networks have shown great potential in medical image segmentation tasks. In this paper, we present a deep convolutional neural network model named Multi-Modal Cardiac Network (MMC-Net) for segmenting three cardiac structures namely right ventricle (RV), left ventricle (LV), and left ventricular myocardium (LVM) from multi-modal CMRI’s. The proposed MMC-Net is designed using a densely connected backbone enabling feature reuse, an atrous convolution module for fusing multi-scale features, and a pixel-classification module for generating the segmentation result. This model was evaluated on a publicly available MS-CMRSeg-2019 challenge dataset in segmentation of RV, LV, and LVM from CMRI scans. The segmentation results from extensive experiments demonstrate our MMC-Net can achieve better segmentation performance compared to other state-of-the-art models, and the existing approaches. Additionally, the generalization ability of the proposed MMC-Net is validated on another publicly available ACDC dataset without fine-tuning. The results demonstrate that the proposed MMC-Net shows a powerful generalisation ability of segmenting RV, LV, and LVM with higher performance.

Background and Objective: LEAD, an environmental toxicant, accounts for 0.6% of the global burden of disease, with the highest burden in developing countries. It can permeate through the foetus blood-brain barrier causing a negative impact on foetal growth and the developing brain. Literacy and awareness related to its impact are low and the clinical establishment for biological monitoring of Blood LEAD Level is low, costly, and time-consuming.Methods: We propose a novel approach to build a low-cost point-of-care analytical device, a screening app that can predict LEAD toxicity levels in maternal blood. A computational model is built that learns from maternal data comprising of blood LEAD level and a set of sociodemographic features, and then predicts the LEAD toxicity level based on the set of input features.Results: Following feature selection methods, the 11-feature set obtained from Boruta algorithm gave the best prediction results. The k-Nearest Neighbour-based model gave 94.00% accuracy whereas the Neural Network-based model gave 92.50% accuracy when tested on 200 participants.Conclusion: The range of features identified in the built models can estimate the underlying function and can accurately model LEAD toxicity prediction and provide an understanding of toxicity level. Early identification and intervention of LEAD-exposed pregnant women will reduce LEAD poisoning in infants and thus, prevent harmful effects on health throughout childhood and adulthood. The built prediction model is beneficial in improving the point of care and hence reducing the cost and the risk involved. It is envisaged that the app will become a part of a screening process to assist healthcare experts at the point of evaluating the LEAD toxicity level in pregnant women. Women screened positive could be given a range of facilities including preliminary counselling to being referred to the health centre for further diagnosis. Steps could be taken to reduce maternal exposure and hence, it could also be possible to mitigate infant’s LEAD exposure by reducing transfer from the pregnant woman.

Event-based vision sensors are a paradigm shift in the way that visual information is obtained and processed. These devices are capable of low-latency transmission of data which represents the scene dynamics. Additionally, low-power benefits make the sensors popular in finite-power scenarios such as high-speed robotics or machine vision applications where latency in visual information is desired to be minimal. The core datatype of such vision sensors is the ‘event’ which is an asynchronous per-pixel signal indicating a change in light intensity at an instance in time corresponding to the spatial location of that sensor on the array. A popular approach to event-based processing is to map events onto a 2D plane over time which is comparable with traditional imaging techniques. However, this paper presents a disruptive approach to event data processing that uses a tree-based filter framework that directly processes raw event data to extract events corresponding to interest point features, which is then combined with a Harris interest point approach to isolate features. We hypothesise that since the tree structure contains the same spatial information as a 2D surface mapping, Harris may be applied directly to the content of the tree, bypassing the need for transformation to the 2D plane. Results illustrate that the proposed approach performs better than other state-of-the-art approaches with limited compromise on the run-time performance.

Building an effective automatic micro expression recognition (MER) system is becoming increasingly desirable in computer vision applications. However, it is also very challenging given the fine-grained nature of the expressions to be recognized. Hence, we investigate if amplifying micro facial muscle movements as a pre-processing phase, by employing Eulerian Video Magnification (EVM), can boost performance of Local Phase Quantization with Three Orthogonal Planes (LPQ-TOP) to achieve improved facial MER across various datasets. In addition, we examine the rate of increase for recognition to determine if it is uniform across datasets using EVM. Ultimately, we classify the extracted features using Support Vector Machines (SVM). We evaluate and compare the performance with various methods on seven different datasets namely CASME, CAS(ME)2, CASME2, SMIC-HS, SMIC-VIS, SMIC-NIR and SAMM. The results obtained demonstrate that EVM can enhance LPQ-TOP to achieve improved recognition accuracy on the majority of the datasets.

This paper presents a novel Reducing-Over-Time (ROT) binary tree structure for event-based vision data and subtypes of the tree structure. A framework is presented using ROT, that takes advantage of the self-balancing and self-pruning nature of the tree structure to extract spatial-temporal information. The ROT framework is paired with an established motion classification technique and performance is evaluated against other state-of-the-art techniques using four datasets. Additionally, the ROT framework as a processing platform is compared with other event-based vision processing platforms in terms of memory usage and is found to be one of the most memory efficient platforms available.

Event-based image processing is a relatively new domain in the field of computer vision. Much research has been carried out on adapting event-based data to comply with established techniques from frame-based computer vision. On the contrary, this paper presents a descriptor which is designed specifically for direct use with event-based data and therefore can be considered to be a pure event-based vision descriptor as it only uses events emitted from event-based vision devices without transforming the data to accommodate frame-based vision techniques. This novel descriptor is known as the Post-stimulus Time-dependent Event Descriptor (P-TED). P-TED is comprised of two features extracted from event data which describe motion and the underlying pattern of transmission respectively. Furthermore a framework is presented which leverages the P-TED descriptor to classify motions within event data. This framework is compared against another state-of-the-art event-based vision descriptor as well as an established frame-based approach.

Autism Spectrum Disorder (ASD) is a developmental disorder that is prevalent globally. Research into detecting autism traditionally focused on behavioural aspects of the condition, however, more recently, focus has shifted to more objective alternatives using techniques such as machine learning and gait analysis. Gait measurements, having been used for person identification, varies from person to person, introducing a lot of intra-subject variance. This applies to the 8 spatial-temporal features used in this study, representing the time that an individual spends in each phase of a gait cycle, collected using a Vicon motion tracking system. The features were averaged across each gait trial that the subjects performed, producing a second set of features with reduced intra-subject variance. Four common classifiers, a Support Vector Machine (SVM), K-Nearest Neighbour (KNN), Random Forests (RF) and a Decision Tree (DT) classifier, were all trained using the two feature sets and their classification rates were compared. The results show that for the RF classifier, reducing the intra-subject variance, was able to successfully increase the classification power. The KNN and DT classifiers experienced a minimal decrease in accuracy, where the SVM suffered the greatest loss when intra-subject variance was reduced. Results overall show that the effect intra-subject variance has on classification power depends heavily on the suitability of the classifier to the initial problem as well as size and class balance of the data.

Neural coding schemes are powerful tools used within neuroscience. This paper introduces three different neural coding scheme formations for event-based vision data which are designed to emulate the neural behaviour exhibited by neurons under stimuli. Presented are phase-of-firing and two sparse neural coding schemes. It is determined that machine learning approaches, i.e. Convolutional Neural Network combined with a Stacked Autoencoder network, produce powerful descriptors of the patterns within events. These coding schemes are deployed in an existing action recognition template and evaluated using two popular event-based data sets.

Gait-based person identification suffers from the problem of different covariate factors such as clothing and carrying objects, which drastically reduce the recognition rate. Most existing methods capture dynamic and static information and remove the covariate factors without any systematic study. However, it has been reported in the literature that the head is one of the important features and the removal of the head from static information decreases the recognition rate. In our preliminary study, we developed a novel random walk (RW)-based gait extraction method that retains the head portion and removes certain static body parts to reduce the effect of covariate factors. The RW-based method is a novel gait feature extraction method and should be exploited more for its discriminative power to separate different body parts efficiently. However, the dynamic part is also significant in gait information, which is not very effectively represented in the RW-based gait extraction method. Therefore, a discrete Fourier transform (DFT)-based frequency component of the gait is considered to represent the dynamic part of gait information. Furthermore, we propose a novel gait recognition algorithm that fuses dynamic and static information from DFT- and RW-based representations. The proposed method systematically retains the discriminative static gait information along with the frequency attribute embedded as the dynamic gait information. Extensive experiments on the Chinese Academy of Sciences, Institute of Automation and the HumanID datasets have been carried out to demonstrate that the proposed fused gait features-based approach outperforms the existing methods, particularly when there are substantial appearance changes.

Children with autism spectrum disorder (ASD) have altered behaviors in communication, social interaction, and activity, out of which communication has been the most prominent disorder among many. Despite the recent technological advances, limited attention has been given to screening and diagnosing ASD by identifying the speech deficiencies (SD) of autistic children at early stages. This research focuses on bridging the gap in ASD screening by developing an automated system to distinguish autistic traits through speech analysis. Data was collected from 40 participants for the initial analysis and recordings were obtained from 17 participants. We considered a three-stage processing system; first stage utilizes thresholding for silence detection and Vocal Activity Detection for vocal isolation, second stage adopts machine learning technique neural network with frequency domain representations in developing a reliant utterance classifier for the isolated vocals and stage three also adopts machine learning technique neural network in recognizing autistic traits in speech patterns of the classified utterances. The results are promising in identifying SD of autistic children with the utterance classifier having 78% accuracy and pattern recognition 72% accuracy.

Research by psychologists have shown that subjects had a preference for a side of a face when it was expressing emotions. This paper seeks to find what accuracies can be attained when only a segment of the face is considered. We show that using one side of the face only reduces accuracy by 0.34% but at half the computationally time required. Various other sections of the face are evaluated for similar performance. We demonstrate that using smaller portions of the face have an expected computation reduction but dont suffer the same degree of accuracy loss. For evaluating we train with a Convolutional Neural Network. To test what portions of a facial image are useful, the full face, half face, eyes, single eye, mouth and half of the mouth are chosen. These images come from the JAFFE, CK+ and KDEF datasets.

Closed-circuit television footage can be used to assemble an often-complex picture of an incident and aid in the identification of suspects after a crime or terrorist attack has occurred. For example, such footage allowed the police to not only identify the 7/7 London bombers but also to piece together the details of the bombers' movements prior to the attack. In the case of the London bombers little attempt was made to disguise their identities but where such identities are concealed it is possible to identify suspects based on other unique biometric characteristics such as the style of walk, referred to as gait. Gait feature–based individual identification has received increased attention from biometrics researchers. In this article, we propose a novel gait biometric methodology that could contribute to the counterterrorism effort and the identification of individuals involved in crime.

Since 9/11, there has been an increased interest in the use of biometric technologies as a way to counter the threat of terrorism and to protect a nation's critical national infrastructure and infrastructure assets. Biometric features can be used to verify an individual's identity or in the identification of an individual. Such features can be divided into two types, namely physical and behavioral characteristics. Physical biometrics are concerned with direct measurements of an individual's body and would include fingerprints, facial geometry and iris patterns. In contrast, behavioral biometrics are concerned with the indirect measurements of an individual, namely their patterns of behavior such as gait, voice and keystroke dynamics. Thus, each individual has a set of unique behavioral biometric features, which they usually adhere to under normal conditions. Based on this concept it is then possible to create different behavioral biometric profiles for each individual that can distinguish one indi...

The use of biometrics has become a popular method to counter the threat of terrorism. Biometric features can be used to confirm identity (biometric authentication) or to identify an individual (biometric identification). Such features include physical (e.g. fingerprints and DNA) and behavioural characteristics (e.g. handwriting and gait). This paper is concerned with one type of behavioural characteristics, namely gait. Gait refers to an individual's style or manner of walk and thus represents one of the body language information signifiers that can identify individuals distinctively in surveillance videos. Recently, gait recognition for individual identification has received increased attention from biometrics researchers as gait can be captured at a distance by using a low-resolution camera. Human gait properties can be affected by clothing and the carrying of objects both of which have been identified within the existing literature as causing difficulties in gait recognition. In this paper, we propose ...

We consider the problem of person identification using gait sequences under normal, carrying bag and different clothing conditions as the main concern. It has been demonstrated that Gait Energy Image (GEI) can attain a better gait recognition rate under normal conditions. However, it has been shown that GEI is not robust enough to handle the carrying bags and different clothing conditions. Instead of GEI, there are several appearance based gait features in the available literature to reduce the effect of covariate factors by keeping dynamic parts and removing the static parts of the gait features under the assumption that the carrying bags and different clothing conditions affect mostly the static parts. It is however shown in the literature that the static parts also contain valuable information and removal of certain static parts such as head by mistake thigh typed certainly decreases the recognition rate.Our main objective has been to increase the gait recognition rate on different clothing and carrying bag covariate gait sequences. Therefore instead of removing static parts, the Joint Sparsity Model (JSM) is applied to identify the carrying bags and different clothings conditions from GEI features. If a set of GEI feature vectors is submitted to JSM model then a common component and an innovations component for each GEI feature are obtained. The innovations component that has unique characteristic to each of features is considered to identify the covariate conditions. The identified covariate conditions are removed from GEI features and a novel gait feature called GEI JSM is generated. The dimension of GEI JSM is reduced using Random Projection (RP) approach and ? 1 -norm minimization technique based sparse representation is used for classification. It is demonstrated that the RP and ? 1 -norm minimization based sparse representation approach provides statistically significant better results than that of the existing individual identification approaches.

A reliable human skin detection method that is adaptable to different human skin colours and illu- mination conditions is essential for better human skin segmentation. Even though different human skin colour detection solutions have been successfully applied, they are prone to false skin detection and are not able to cope with the variety of human skin colours across different ethnic. Moreover, existing methods require high computational cost. In this paper, we propose a novel human skin de- tection approach that combines a smoothed 2D histogram and Gaussian model, for automatic human skin detection in colour image(s). In our approach an eye detector is used to refine the skin model for a specific person. The proposed approach reduces computational costs as no training is required; and it improves the accuracy of skin detection despite wide variation in ethnicity and illumination. To the best of our knowledge, this is the first method to employ fusion strategy for this purpose. Qualitative and quantitative results on three standard public datasets and a comparison with state-of-the-art methods have shown the effectiveness and robustness of the proposed approach., Accepted in IEEE Transactions on Industrial Informatics, vol. 8(1), pp. 138-147, new skin detection + ground truth (Pratheepan) dataset

Many silhouettes based features are proposed for gait recognition. But these methods suffer with covariate factors such as clothing and carrying objects. These covariate factors mostly attached with silhouettes and make gait recognition much difficult. Therefore we proposed a new silhouette based feature to analyse the influence of body parts and increase the recognition rate for individual identification.

The Internet is a widely open source to everyone to access Web pages. Using a web browser anyone can access websites. Because of this facility people can easily download images from websites without the owner's knowledge and use them in their own documents. Also image content may be modified for illegal purposes. Therefore a system is needed to authenticate images over the Web. Web image authentication is a challenging task that requires web crawlers to track and download images for authentication. Most of the known web image tracking engines such as Tin Eye and PicScout retrieve images according to the image infringement of the original image. However, these systems do not have the facility to authenticate the retrieved image, i.e. whether the retrieved image is similar to the original image or any image content alteration has occurred in the retrieved image and who is the copyrighted owner of the retrieved image. In order to solve the above mentioned drawbacks this paper presents a framework to protect image content, track it over the internet and authenticate the content. The proposed framework is based on self-embedding (i.e. where secret data and a binary version of the image are encrypted and embedded into the image), tracking (i.e. where a web crawler traverses over the internet to download images) and self-authentication (i.e. where the binary version of the hidden data is extracted to authenticate the image). Also another advantage of the proposed system is that it does not need the original image for the authentication process.