SciTePress - Affiliation Profile

Affiliation

Select a Subject
Refine Ontology Subject

Faculty of Physiotherapy at University 21, University of Applied Sciences, Germany

Context (please use complete words only)

Field

Sort by

24 Papers

Uncertainty and Feature-Based Weighted Loss for 3D Wheat Part Segmentation

Topic: 3D Deep Learning

Authors: R. Reena ¹ , John Doonan ² , Kevin Williams ² , Fiona Corke ² , Huaizhong Zhang ¹ and Yonghuai Liu ¹

Affiliations: ¹ Edge Hill University, Lancashire, U.K. , ² National Plant Phenomics Centre, Aberystwyth University, U.K.

Digital Object Identifier: 10.5220/0013312300003912

Proceedings of VISAPP 2025, ISBN: 978-989-758-728-3

Quick Abstract

Deep learning techniques and point clouds have proved their efficacy in 3D segmentation tasks of objects. Nevertheless, the accurate plant organ segmentation is a formidable challenge due to their complex structure and variability. Furthermore, presence of over-represented and under-represented parts, occlusion, and uneven distribution complicates the 3D part segmentation tasks. Even though deep learning techniques often exhibit exceptional performance, they also face challenges in applications where accurate trait estimation is required. To handle these issues, we propose a novel uncertainty and feature based weighted loss that incorporates uncertainty metrics and features of the plant or crop. We use Gradient Attention Module (GAM) with PointNet++ baseline to validate our approach. By dynamically introducing uncertainty and feature scores into the training process, it promotes more balanced learning. Through comprehensive evaluation, we illustrate the advantages of UFL (Uncertainty and Feature based Loss) as compared to standard CE (Cross entropy loss) with our own constructed real Wheat dataset. The outcomes demonstrate consistent improvements in Accuracy (ranging from 0.9% to 4.2%) and Ear mIoU (ranging from 1.8% to 15.3%) over the standard Cross-Entropy (CE) loss function. As a result, our work contributes to the development of more robust and reliable segmentation models. This approach not only pushes forward the boundaries of precision agriculture but also has the potential to influence related areas where accurate segmentation is pivotal.

A Tailored Internet of Things Lighting Solution to Support Circadian Rhythms and Wellbeing for People Living with Dementia

Topics: Smart Environments and Housing; Remote Monitoring; Autonomy and Active Ageing; Psychological Dimension of Aging Population; Human Communication and Behavioural Studies of Senior Citizens; Ambient Technology for Daily Living; Ambient Technology for Health, Emergency and Care; Internet of Things and Smart Devices for Independent Living

Authors: Kate Turley ¹ , Joseph Rafferty ¹ , Raymond Bond ¹ , Assumpta Ryan ² , Maurice Mulvenna ¹ and Lloyd Crawford ³

Affiliations: ¹ School of Computing, Ulster University, 2-24 York St., Belfast, N. Ireland , ² School of Nursing and Pandemic Science, Ulster University, Northland Rd, Derry, N. Ireland , ³ Chroma Lighting, 213-215 Donegal Rd, Belfast, N. Ireland

Digital Object Identifier: 10.5220/0012650100003699

Proceedings of ICT4AWE 2024, ISBN: 978-989-758-700-9

Quick Abstract

Light is a requirement for setting and maintaining the body’s circadian rhythm, however our knowledge of the spectral content, timing and duration of lighting exposure for the indoors is not well defined. For people living with dementia, this knowledge gap is important to address since they experience more heavily disrupted circadian rhythms, which can heighten symptoms of sundowning, agitation, low mood and poor sleep quality. This paper focuses on the required design aspects for a dynamic lighting and sensing device tailored towards supporting the wellbeing of people living with dementia. The authors discuss the current understanding of lighting for health, identify the gaps to be addressed and propose the design and research protocol for an indoor lighting and sensing solution. The device is currently deployed within a care home and analysis of results is forthcoming.

Feature Extraction Methods for Neural Networks in the Classification of Structural Health Anomalies

Topics: Computational Neuroscience; Complex Artificial Neural Network Based Systems and Dynamics; Applications: Image Processing and Artificial Vision, Pattern Recognition, Decision Making, Industrial and Real World Applications, Financial Applications, Neural Prostheses and Medical Applications, Neural Based Data Mining and Complex Information Process; Bio-inspired models and solutions

Authors: Natasha Hamilton ¹ , Jim Harkin ¹ , Liam McDaid ¹ , Junxiu Liu ¹ and Eoghan Furey ²

Affiliations: ¹ School of Computing, Engineering and Intelligent Systems, Ulster University, Northland Road, Derry, U.K. , ² Department of Computing, Atlantic Technological University, Port Road, Donegal, Ireland

Digital Object Identifier: 10.5220/0012184800003595

Proceedings of NCTA 2023, ISBN: 978-989-758-674-3

Quick Abstract

Failure of large complex structures such as buildings and bridges can have monumental repercussions such as human mortality, environmental destruction and economic consequences. It is therefore paramount that detection of structural damage or anomalies are identified and managed early. This highlights the need to develop automated Structural Health Monitoring (SHM) systems that can continuously allow the safety status of structures to be determined, even in the worst and most isolated conditions, to ultimately help prevent destruction and save lives. Signal processing is a crucial step to detecting structural anomalies and recent work demonstrates the opportunities for neural networks, however the encoding of data for SHM requires the extraction of features due to often, noisy data. This paper focuses on feature extraction methods for artificial neural networks (ANNs) and spiking neural networks (SNNs) and aims to identify bespoke features which enable SNNs to encode data and perform the classification of anomalies. Results show that extraction of particular features in large real-world applications improve the classification accuracy of SNNs.

Evaluation of a System for Named Entity Recognition in a Knowledge Management Ecosystem

Topics: Natural Language Processing; Applications and Case-Studies

Authors: Philippe Tamla ¹ , Florian Freund ¹ , Matthias Hemmje ¹ and Paul Mc Mc Kevitt ²

Affiliations: ¹ University of Hagen, Faculty of Mathematics and Computer Science, 58097 Hagen, Germany , ² Ulster University, Faculty of Arts, Humanities and Social Sciences, Londonderry BT48 7JL, North Ireland

Digital Object Identifier: 10.5220/0011374000003335

Proceedings of KEOD 2022/IC3K, ISBN: 978-989-758-614-9

Quick Abstract

In this research paper, the evaluation of a new system for machine learning-based Named Entity Recognition is presented. After introducing our approach supporting two fundamental tasks for training Named Entity Recognition models using machine learning (data cleanup and data annotation), our features, prototype and evaluation methodologies are described. Also, the results of our performed quantitative and qualitative experiments validating our approach and user interface are shown. Finally, our evaluation results are discussed to derive challenges for future work.

Evaluation of Generative Adversarial Network Generated Super Resolution Images for Micro Expression Recognition

Topics: Feature Selection and Extraction; Image and Video Analysis and Understanding

Authors: Pratikshya Sharma ¹ , Sonya Coleman ¹ , Pratheepan Yogarajah ¹ , Laurence Taggart ² and Pradeepa Samarasinghe ³

Affiliations: ¹ School of Computing, Engineering & Intelligent Systems, Ulster University, Northern Ireland, U.K. , ² School of Nursing & Health Research, Ulster University, Northern Ireland, U.K. , ³ Department of Information Technology, Malabe, Sri Lanka

Digital Object Identifier: 10.5220/0010820100003122

Proceedings of ICPRAM 2022, ISBN: 978-989-758-549-4

Quick Abstract

The Advancements in micro expression recognition techniques are accelerating at an exceptional rate in recent years. Envisaging a real environment, the recordings captured in our everyday life are prime sources for many studies, but these data often suffer from poor quality. Consequently, this has opened up a new research direction involving low resolution micro expression images. Identifying a particular class of micro expression among several classes is extremely challenging due to less distinct inter-class discriminative features. Low resolution of such images further diminishes the discriminative power of micro facial features. Undoubtedly, this increases the recognition challenge by twofold. To address the issue of low-resolution for facial micro expression, this work proposes a novel approach that employs a super resolution technique using Generative Adversarial Network and its variant. Additionally, Local Binary Pattern & Local phase quantization on three orthogonal planes are used for extracting facial micro features. The overall performance is evaluated based on recognition accuracy obtained using a support vector machine. Also, image quality metrics are used for evaluating reconstruction performance. Low resolution images simulated from the SMIC-HS dataset are used for testing the proposed approach and experimental results demonstrate its usefulness.

Fast Human Activity Recognition

Topics: Video Surveillance and Event Detection; Motion Detection; Action Recognition; Deep Learning and Neural Networks; Classification and Clustering; Image Analysis and Scene Understanding; Video Understanding; Video Indexing and Annotation; High-speed Computer Vision; Crowds Analysis

Authors: Shane Reid ¹ , Sonya Coleman ¹ , Dermot Kerr ¹ , Philip Vance ¹ and Siobhan O’Neill ²

Affiliations: ¹ School of Computing, Engineering and Intelligent Systems, Ulster University Magee Campus, Derry/Londonderry, Northern Ireland , ² School of Psychology, Ulster University Coleraine Campus, Coleraine, Northern Ireland

Digital Object Identifier: 10.5220/0010420300910098

Proceedings of IMPROVE 2021, ISBN: 978-989-758-511-1

Quick Abstract

Human activity recognition has been an open problem in computer vision for almost two decades. In that time there have been many approaches proposed to solve this problem, but very few have managed to solve it in a way that is sufficiently computationally efficient for real time applications. Recently this has changed, with keypoint based methods demonstrating a high degree of accuracy with low computational cost. These approaches take a given image and return a set of joint locations for each individual within an image. In order to achieve real time performance, a sparse representation of these features over a given time frame is required for classification. Previous methods have achieved this by using a reduced number of keypoints, but this approach gives a less robust representation of the individual’s body pose and may limit the types of activity that can be detected. We present a novel method for reducing the size of the feature set, by calculating the Euclidian distance and the direction of keypoint changes across a number of frames. This allows for a meaningful representation of the individuals movements over time. We show that this method achieves accuracy on par with current state of the art methods, while demonstrating real time performance.

Augmenting Guideline-based CDSS with Experts’ Knowledge

Topics: Knowledge Management; Decision Support Systems; Practice-based Research Methods for Healthcare IT

Authors: Nekane Larburu ¹ , Naiara Muro ¹ , Iván Macía ¹ , Eider Sánchez ² , Hui Wang ³ , John Winder ³ , Jacques Boaud ⁴ and Brigitte Séroussi ⁵

Affiliations: ¹ Vicomtech-IK4 and Biodonostia, Spain , ² NARU, Spain , ³ Ulster University, United Kingdom , ⁴ AP-HP, DRCD, Sorbonne Universités, UPMC Univ. Paris 06, INSERM and Université Paris 13, France , ⁵ Sorbonne Universités, UPMC Univ. Paris 06, INSERM, Université Paris 13, AP-HP, Hôpital Tenon and DSP, France

Digital Object Identifier: 10.5220/0006213903700376

Proceedings of HEALTHINF 2017/BIOSTEC, ISBN: 978-989-758-213-4

Quick Abstract

Over the past years, clinical guidelines have increasingly become part of the clinical daily practice in order to provide best available Evidence-Based-Medicine services. Hence, their formalization as computer interpretable guidelines (CIG) and their implementation in clinical decision support systems (CDSSs) are emerging to support clinicians in their decision making process and potentially improve medical outcomes. However, guideline compliancy in the clinical daily practice is still “low”. Some of the reasons for such low compliance rate are (i) lack of a complete guideline to cover special clinical cases (e.g. oncogeriatric cases), (ii) absence of parameters that current guidelines do not consider (e.g. lifestyle) and (iii) absence of up-to-date guidelines due to lengthy validation procedures. In this paper we present a novel method to build a CDSS that, besides integrating CIGs, stores experts’ knowledge to enrich the CDSS and provide best support to clinicians. The knowledge includes new evidence collected over time by the systematic usage of CDSSs.

Related Ontology Subjects/Areas/Topics (19)

Artificial Intelligence	Decision Support Systems, Remote Data Analysis	Knowledge-Based Systems
Biomedical Engineering	Enterprise Information Systems	Ontologies and the Semantic Web
Business Analytics	Health Engineering and Technology Applications	Practice-based Research Methods for Healthcare IT
Cardiovascular Technologies	Health Information Systems	Society, e-Business and e-Government
Computing and Telecommunications in Cardiology	Information Systems Analysis and Specification	Symbolic Systems
Data Engineering	Knowledge Management	Web Information Systems and Technologies
Decision Support Systems

Si elegans: A Computational Model of C. elegans Muscle Response to Light

Authors: Alicia Costalago Meruelo , Pedro Machado , Kofi Appiah and T. Martin McGinnity

Affiliation: Nottingham Trent University, United Kingdom

Digital Object Identifier: 10.5220/0005712201210126

Proceedings of NeBICA 2015/NEUROTECHNIX, ISBN: 978-989-758-161-8

Quick Abstract

It has long been the goal of computational neuroscientists to understand animal nervous systems, but their vast complexity has made it very difficult to fully understand even basic functions such as movement. The C. elegans nematode offers the opportunity to study a fully described connectome and link neural network to behaviour. In this paper a model of the responses of the body wall muscle in C. elegans to a random light stimulus is presented. An algorithm has been developed that tracks synapses in the nematode nervous system from the stimulus in the phototaxis sensory neurons to the muscles cells. A linear second order model was used to calculate the isometric force in each of the C. elegans body wall muscle cells. The isometric force calculated resembles that of previous investigations in muscle modelling.

Exploring Neural Principles with Si elegans, a Neuromimetic Representation of the Nematode Caenorhabditis elegans

Authors: Axel Blau ¹ , Frank Callaly ² , Seamus Cawley ² , Aedan Coffey ² , Alessandro de Mauro ³ , Gorka Epelde ³ , Lorenzo Ferrara ¹ , Finn Krewer ² , Carlo Liberale ¹ , Pedro Machado ⁴ , Gregory Maclair ³ , Thomas Martin McGinnity ⁴ , Fearghal Morgan ² , Andoni Mujika ³ , Alexey Petrushin ¹ , Gautier Robin ³ and John Wade ⁴

Affiliations: ¹ Fondazione Istituto Italiano di Tecnologia (IIT), Italy , ² National University of Ireland, Ireland , ³ Vicomtech-IK4, Spain , ⁴ University of Ulster and Nottingham Trent University, United Kingdom

Digital Object Identifier: 10.5220/0005190701890194

Proceedings of NeBICA 2014/NEUROTECHNIX, ISBN: 978-989-758-056-7

Quick Abstract

Biological neural systems are powerful, robust and highly adaptive computational entities that outperform conventional computers in almost all aspects of sensory-motor integration. Despite dramatic progress in information technology, there is a big performance discrepancy between artificial computational systems and brains in seemingly simple orientation and navigation tasks. In fact, no system exists that can faithfully reproduce the rich behavioural repertoire of the tiny worm Caenorhabditis elegans which features one of the simplest nervous systems in nature made of 302 neurons and about 8000 connections. The Si elegans project aims at providing this missing link. This article is sketching out the main platform components.

Si elegans - Computational Modelling of C. elegans Nematode Nervous System using FPGAs

Authors: Pedro Machado ¹ , John Wade ¹ and T. M. Mcginnity ²

Affiliations: ¹ University of Ulster, Ireland , ² University of Ulster and Nottingham Trent University, United Kingdom

Digital Object Identifier: 10.5220/0005169301690176

Proceedings of NeBICA 2014/NEUROTECHNIX, ISBN: 978-989-758-056-7

Quick Abstract

It has long been the goal of computational neuroscientists to understand and harness the parallel computational power of the mammalian nervous system. However, the vast complexity of such a nervous system has made it very difficult to fully understand even the most basic of functions such as movement and learning and accordingly there has been increasing attention paid to the development of emulations of simpler systems. One such system is the C. elegans nematode, which has been widely studied in recent years and there now exists a vast wealth of biological knowledge about its nervous structure, function and connectivity. The Si elegans EU FP7 project aims to develop a Hardware Neural Network (HNN) to accurately replicate the C. elegans nervous system behaviour to enable neuroscientists to better understand these basic functions. To fully replicate the C. elegans biological system requires powerful computing technologies, based on parallel processing, for real-time computation and therefore will use Field Programmable Gate Arrays (FPGAs) to achieve this. In this paper an overview of the complete hardware system required to fully realise Si elegans is presented along with an early small scale implementation of the hardware system.

Context (please use complete words only)

Field

Sort by

Topic(s):

Filter listed results, by

Publication date

From:

To:

Author

Locate Author in list below

Data pager
Data pager
1 2 3 4 5 6 7 8 9 10 ... Page 1 of 11, items 1 to 10 of 102.

Add Author(s) to filter:

Martin M. Mcginnity (7)

S. Coleman (7)

Dermot Kerr (4)

Maurice Mulvenna (3)

Qingxiang Wu (3)

P. Machado (3)

John Wade (2)

Jim Harkin (2)

Fiona M. K. Corke (2)

Paul Kevitt (2)

Conference

Locate Conference in list below

Data pager
Data pager
1 2 Page 1 of 2, items 1 to 10 of 17.

Add Conference(s) to filter:

NCTA(4)

NEUROTECHNIX(3)

KDIR(2)

DCNET(2)

ECTA(1)

FCTA(1)

KEOD(1)

PECCS(1)

SENSORNETS(1)

VISAPP(1)

Title:
ISBN:
Year:
Acronym:
Subject:

Title:
Author:
Affiliation:
Subject:

Name:
Affiliation:
Country:
Conference:
Subject:

Name:
Country:
Subject: