Tomas Henrique Maul

Tomas Henrique Maul
University of Nottingham, Malaysia Campus | nottingham · School of Computer Science

PhD

About

69
Publications
22,016
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
218
Citations
Citations since 2016
35 Research Items
174 Citations
2016201720182019202020212022010203040
2016201720182019202020212022010203040
2016201720182019202020212022010203040
2016201720182019202020212022010203040
Introduction
I’m interested in both biological and artificial neural computation. In the former domain I have focused primarily on creating biologically plausible models of visual functions. In the latter domain I have focused primarily on higher-order neural networks and on using evolutionary algorithms for the optimization of new types of hybrid artificial neural networks (HANN). I am currently studying the metaphor of neuronal diversity (as seen for instance in the retina) and its implications to HANN.
Additional affiliations
January 2009 - present
University of Nottingham, Malaysia Campus
Position
  • Professor (Assistant)
Education
June 2002 - January 2006
University of Malaya
Field of study
September 2000 - September 2001
Imperial College of Science, Technology & Medicine, University of London
Field of study
September 1999 - September 2000
Imperial College of Science, Technology & Medicine, University of London.
Field of study

Publications

Publications (69)
Article
Automatic extraction of distinctive features from a visual information stream is challenging due to the large amount of information contained in most image data. In recent years deep neural networks (DNNs) have gained outstanding popularity for solving visual information processing tasks. This study reports novel contributions, including a new DNN...
Article
Full-text available
This paper is concerned with the problem of improving the convergence properties of evolutionary neural networks, particularly in the context of hybrid neural networks that adopt a diversity of transfer functions at their nodes, i.e.: neural diversity machines. The paper explores the potential of solution complementarity, in the context of pattern...
Conference Paper
Full-text available
Artificial Neural Networks are robust in their appli-cations; however, choosing the appropriate neural network model best suited for a particular problem is usually just a case of trial and error. In this paper, we present self-adaptive learning for Artificial Neural Networks as a direction towards efficient learning. Our hypothesis is that the neu...
Article
The current paper introduces the concept of neural diversity machines (NDM) which, refers to hybrid artificial neural networks (HANN) with conditions on the minimum number of functions available to the network, amongst several other properties. The paper demonstrates how NDM networks can be optimized for solving different problems. The results demo...
Conference Paper
Full-text available
In this paper we report an interesting observation pertaining to denoising based on the optimization of image processing chains. Although often a goal in itself, denoising is usually performed in order to minimize the detrimental effects of noise in the subsequent stages of an algorithm. Typically, denoising is carried out as an early pre-processin...
Article
Full-text available
Despite the promising results of deep learning research, construction industry applications are still limited. Facility Management (FM) in construction has yet to take full advantage of the efficiency of deep learning techniques in daily operations and maintenance. Heating, Ventilation, and Air Conditioning (HVAC) is a major part of Facility Manage...
Article
Full-text available
Water pollution is a widespread problem, with lakes, rivers, and oceans contaminated by an increasing amount of microplastics and other pollutants. Microplastic counting from microscope images is a laborious, time-consuming, and error-prone task. The ability of researchers to automate the detection and counting of microplastics would accelerate res...
Article
Full-text available
Anomaly detection is referred to a process in which the aim is to detect data points that follow a different pattern from the majority of data points. With the rapid development of computer technology, protecting networks from various threats such as network intruders is becoming crucial. Traditional anomaly detection methods suffer from several we...
Chapter
Deep neural networks tend to be accurate but computationally expensive, whereas ensembles tend to be fast but do not capitalize on hierarchical representations. This paper proposes an approach that attempts to combine the advantages of both approaches. Hierarchical ensembles represent an effort in this direction, however they are not compositional...
Article
Full-text available
The combination and aggregation of knowledge from multiple neural networks can be commonly seen in the form of mixtures of experts. However, such combinations are usually done using networks trained on the same tasks, with little mention of the combination of heterogeneous pre-trained networks, especially in the data-free regime. The problem of com...
Article
After becoming independent in 1957, Malaysia continued as an agricultural country but quickly grew into a manufacturing nation in a relatively short time. Literally from nowhere, the manufacturing sector now commands more than 38% of the nation’s GDP overtaking the agriculture sector which commands just slightly above 7%. In addition to the multina...
Article
Full-text available
A high dimensional low sample size (HDLSS) dataset typically contains many features but a limited number of samples. It is commonly found in domains such as microarray data and medical imaging. When sample size is small, the population probability density function (PDF) of a HDLSS dataset may not be well represented, causing difficulties of applyin...
Preprint
Full-text available
The growing capacity of neural networks has strongly contributed to their success at complex machine learning tasks and the computational demand of such large models has, in turn, stimulated a significant improvement in the hardware necessary to accelerate their computations. However, models with high latency aren't suitable for limited-resource en...
Preprint
Full-text available
The combination and aggregation of knowledge from multiple neural networks can be commonly seen in the form of mixtures of experts. However, such combinations are usually done using networks trained on the same tasks, with little mention of the combination of heterogeneous pre-trained networks, especially in the data-free regime. This paper propose...
Article
Full-text available
Capsule network (CapsNet) was introduced as an enhancement over convolutional neural networks, supplementing the latter’s invariance properties with equivariance through pose estimation. CapsNet achieved a very decent performance with a shallow architecture and a significant reduction in parameters count. However, the width of the first layer in Ca...
Conference Paper
Full-text available
Word embeddings are numerical distributed word representations that have recently sparked significant interest in the research community. They are used by Google and Facebook corporations in various applications such as enhancing search and recommendation engines. Many word embedding techniques such as word2vec, doc2vec, glove, BERT, RoBERT and oth...
Chapter
Automatic breast cancer classification benefits pathologists in obtaining fast and precise diagnoses and improving early detection. However, the performance of deep learning models depends greatly on the quality and quantity of the datasets used. Due to the complexity and high costs of patient data collection, many medical datasets, particularly fo...
Preprint
This paper proposes an autoencoder (AE) that is used for improving the performance of once-class classifiers for the purpose of detecting anomalies. Traditional one-class classifiers (OCCs) perform poorly under certain conditions such as high-dimensionality and sparsity. Also, the size of the training set plays an important role on the performance...
Preprint
Full-text available
Anomaly detection is referred to as a process in which the aim is to detect data points that follow a different pattern from the majority of data points. Anomaly detection methods suffer from several well-known challenges that hinder their performance such as high dimensionality. Autoencoders are unsupervised neural networks that have been used for...
Preprint
Full-text available
This paper proposes an autoencoder (AE) that is used for improving the performance of once-class clas-sifiers for the purpose of detecting anomalies. Traditional one-class classifiers (OCCs) perform poorly under certain conditions such as high-dimensionality and sparsity. Also, the size of the training set plays an important role on the performance...
Preprint
Lifelong learning is a very important step toward realizing robust autonomous artificial agents. Neural networks are the main engine of deep learning, which is the current state-of-the-art technique in formulating adaptive artificial intelligent systems. However, neural networks suffer from catastrophic forgetting when stressed with the challenge o...
Preprint
Full-text available
Outlier detection (also known as anomaly detection or deviation detection) is a process of detecting data points in which their patterns deviate significantly from others. It is common to have outliers in industry applications, which could be generated by different causes such as human error, fraudulent activities, or system failure. Recently, dens...
Article
Full-text available
Data augmentation is a well known technique that is frequently used in machine learning tasks to increase the number of training instances and hence decrease model over-fitting. In this paper we propose a data augmentation technique that can further boost the performance of satellite image super resolution tasks. A super-resolution convolutional ne...
Article
Full-text available
Recent progress in computer vision has pushed the limit of facial recognition from human identification to micro-expressions (MEs). However, the visual analysis of MEs is still a very challenging task because of the short occurrence and insignificant intensity of the underlying signals. To date, the accuracy of recognizing hidden emotions from fram...
Article
Full-text available
Artificial neural networks (ANNs) have achieved significant success in tackling classical and modern machine learning problems. As learning problems grow in scale and complexity, and expand into multi-disciplinary territory, a more modular approach for scaling ANNs will be needed. Modular neural networks (MNNs) are neural networks that embody the c...
Preprint
Full-text available
Convolutional neural networks (CNNs) are known for their good performance and generalization in vision-related tasks and have become state-of-the-art in both application and research-based domains. However, just like other neural network models, they suffer from a susceptibility to noise and adversarial attacks. An adversarial defence aims at reduc...
Preprint
Capsule network (CapsNet) was introduced as an enhancement over convolutional neural networks, supplementing the latter's invariance properties with equivariance through pose estimation. CapsNet achieved a very decent performance with a shallow architecture and a significant reduction in parameters count. However, the width of the first layer in Ca...
Article
Full-text available
How to design and train increasingly large neural network models is a topic that has been actively researched for several years. However, while there exists a large number of studies on training deeper and/or wider models, there is relatively little systematic research particularly on the effective usage of wide modular neural networks. Addressing...
Article
Current satellite remote-sensing technologies enable the timely and detailed monitoring of human activities on the Earth’s surface. Sub-metre spatial resolution satellite images can picture the ongoing works of railway and highway construction. The synoptic view of satellite images is useful to assist the monitoring and management of such construct...
Conference Paper
Full-text available
Crop that are currently underutilized can play a major role in diversifying food sources and combating climate variability. One major obstacle for wider adoption of these species is the lack of information on the geographic areas where these crops are currently grown. These crops are typically grown in marginal lands through subsistence agriculture...
Article
Full-text available
One of the common problems of neural networks, especially those with many layers, consists of their lengthy training time. We attempt to solve this problem at the algorithmic level, proposing a simple parallel design which is inspired by the parallel circuits found in the human retina. To avoid large matrix calculations, we split the original netwo...
Article
Full-text available
The inconsistency between the freely available remote sensing datasets and crowd-sourced data from the resolution perspective forms a big challenge in the context of data fusion. In classical classification problems, crowd-sourced data are represented as points that may or not be located within the same pixel. This discrepancy can result in having...
Conference Paper
Full-text available
In an attempt to solve the lengthy training times of neural networks, we proposed Parallel Circuits (PCs), a biologically inspired architecture. Previous work has shown that this approach fails to maintain generalization performance in spite of achieving sharp speed gains. To address this issue, and motivated by the way Dropout prevents node co-ada...
Article
Full-text available
Nowadays, an advanced remote sensing technology acquires huge amounts of Earth surface details with multi-spatial, multi-temporal and multi-spectral resolutions that have changed drastically the size and structure of data. In order to process such big remote sensing data, it is vital to adopt a new approach, i.e. Cloud Computing, which is an elasti...
Article
Full-text available
Loss of vision is a severe impairment to the dominant sensory system. It often has a catastrophic effect upon the sufferer, with knock-on effects to their standard of living, their ability to support themselves, and their care-givers lives. Research into visual impairments is multi-faceted, focusing on the causes of these debilitating conditions as...
Article
Full-text available
Loss of vision is a severe impairment to the dominant sensory system. It often has a catastrophic effect upon the sufferer, with knock-on effects to their standard of living, their ability to support themselves, and their care-givers lives. Research into visual impairments is multi-faceted, focusing on the causes of these debilitating conditions as...
Article
Full-text available
Loss of vision is a severe impairment to the dominant sensory system. It often has a catastrophic effect upon the sufferer, with knock-on effects to their standard of living, their ability to support themselves, and their care-givers lives. Research into visual impairments is multi-faceted, focusing on the causes of these debilitating conditions as...
Article
Full-text available
In this paper, we report an interesting observation pertaining to new image processing pipeline for membrane detection suggested by optimization experiments. Denoising is usually performed in order to minimize the detrimental effects that noise has on the subsequent stages of an algorithm. Thus Denoising is typically carried out as an early pre-pro...
Conference Paper
Full-text available
Although there have been a few approaches to achieve the goal of fault tolerance by diversifying redundancy of the individual networks that make up a neural network ensemble, some of which include ensembles of neural networks of different sizes, and ensembles of different models of neural networks such as Radial Basis Function Networks and Multilay...
Conference Paper
Full-text available
This paper is concerned with the problem of optimizing deep neural networks with diverse transfer functions using evolutionary methods. Standard evolutionary (SEDeeNN) and cooperative coevolutionary methods (CoDeeNN) were applied to three different architectures characterized by different constraints on neural diversity. It was found that (1) SEDee...
Conference Paper
Full-text available
Recent research has demonstrated the great capability of deep belief networks for solving a variety of visual recognition tasks. However, primary focus has been on modelling higher level visual features and later stages of visual processing found in the brain. Lower level processes such as those found in the retina have gone ignored. In this paper,...
Conference Paper
Full-text available
Local Contrast Hole Filling Algorithm for Neural Slices Membrane Detection (LCHF) algorithm is non-learning, simple, easily adopted, and undependable on ground-truth; and it can recognize membrane and eliminates organelles, using a very simple algorithm that consist of short sequences of basic processing steps yet can be relatively competitive. Her...
Article
Full-text available
Problem signatures are patterns that reveal a glimpse of the computational strategy most likely to be suitable for a given problem. Such a pattern could be the preferred choice of the activation and output functions for a given problem in neural networks that implement transfer functions optimization. We refer to these patterns as first-order signa...
Conference Paper
Fluorescent lamps are becoming popular as an indoor light source in many commercial sites. While high speed video cameras can detect variations in the illumination levels from such light sources, they can be quite costly to operate. Moreover, most popular security video cameras can only operate up to a maximum speed of 30 frames per second. Would t...
Conference Paper
This paper is a follow-up study on an earlier introductory paper on Neural Diversity Machines (NDM). NDMs are a subclass of Hybrid Artificial Neural Networks (HANNs), which are digital representations of biological neural networks present in the human brain. As opposed to traditional artificial neural networks (ANNs) which tend to be focused around...
Conference Paper
Full-text available
In this paper we report an interesting observation pertaining to denoising based on the optimization of image processing chains. Although often a goal in itself, denoising is usually performed in order to minimize the detrimental effects of noise in the subsequent stages of an algorithm. Typically, denoising is carried out as an early pre-processin...
Conference Paper
Full-text available
Transfer functions play an important role in artificial neural networks. They enable neural networks to form decision boundaries of different shapes and forms. However, transfer function optimisation has received relatively little research attention. In this paper, we present an approach for training neural networks that use transfer functions pool...
Article
Full-text available
Visual to auditory conversion systems have been in existence for several decades. Besides being among the front runners in providing visual capabilities to blind users, the auditory cues generated from image sonification systems are still easier to learn and adapt to compared to other similar techniques. Other advantages include low cost, easy cust...
Article
The electrode resolution of current retinal prostheses is still far from matching the densities of retinal neurons. Decreasing electrode diameter increases impedance levels thus deterring effective stimulation of neurons. One solution is to increase the surface roughness of electrodes, which can be done via nanoparticle coatings. This paper explore...
Conference Paper
Full-text available
The Hermann grid was first described and discussed by the physiologist Ludimar Hermann in 1870. It is composed of white horizontal and vertical bars on a black background [1]. Subjects perceive black or gray smudges at the intersections of white bars when looking at the grid. This effect was discussed by Baumgartner who proposed a theory related to...
Conference Paper
Full-text available
This paper describes a simulation model of nanoparticle assemblies formed by spray deposition. The simulation deposits nanoparticles via a semi-mechanistic process, which in spite of its simplicity generates morphologies of considerable complexity. The experiments reveal several relationships between nanoparticle parameters such as their relative p...
Article
Full-text available
The retina still poses many structural and computational questions. Structurally, for example, it is not yet clear how many distinct horizontal cell (HC) types the primate retina contains and what the exact patterns of connections between photoreceptors (PRs) and HCs consist of. Computationally, it is not yet clear, for instance, what functions are...
Article
This paper demonstrates how unsupervised learning based on Hebb-like mechanisms is sufficient for training second-order neural networks to perform different types of motion analysis. The paper studies the convergence properties of the network in several conditions, including different levels of noise and motion coherence and different network confi...
Conference Paper
Full-text available
Age-related macular degeneration and retinitis pigmentosa are two of the most common diseases that cause degeneration in the outer retina, which can lead to several visual impairments up to blindness. Vision restoration is an important goal for which several different research approaches are currently being pursued. We are concerned with restoratio...
Conference Paper
Luminophonics is a system that aims to maximize cross-modality conversion of information, specifically from the visual to auditory modalities, with the motivation to develop a better assistive technology for the visually impaired by using image sonification techniques. The project aims to research and develop generic and highly-configurable compone...
Conference Paper
Full-text available
In this paper a new model of the Outer Plexiform Layer (OPL) of the human retina is presented. The model, which is a multi-resolution Linear Recurrent Neural Net- work (LRNN) defined by 31 parameters, was subjected to several optimization experiments targeting different low-level visual functions involving the control of noise, brightness, contrast...
Conference Paper
Ever since the beginning of research on gait recognition, the main focus has been on investigating gait as a biometric. For security surveillance systems, the detection of suspicious behavior is considered to be more relevant than the recognition of identity, which applies more to security authentication systems. Hence, this research uses gait feat...
Article
Full-text available
This paper is concerned with the problem of visual pose estimation, which entails, for example, the estimation of object translations. It adopts a correspondence based approach in general, and in particular, looks into a neural network implementation of the approach. The objective of the paper is to demonstrate how the approach can be learnt via th...
Patent
A method to detect an abnormal event in a surveillance system using wide view video images is disclosed herein. More particularly the invention provides a solution to overcome the image distortions that are associated with the wide video images.
Conference Paper
There has been a significant drop in the cost as well as an increase in the quality of imaging sensors due to stiff competition as well as production improvements. Consequently, real-time surveillance of private or public spaces which relies on such equipment is gaining wider acceptance. While the human brain is very good at image analysis, fatigue...