arXiv subjects

Kok Yew Ng

Publications and source records attributed to Kok Yew Ng.

6 recordsLinked to original sources

arXiv2025

Design and Development of a Robust Tolerance Optimisation Framework for Automated Optical Inspection in Semiconductor Manufacturing

Automated Optical Inspection (AOI) is widely used across various industries, including surface mount technology in semiconductor manufacturing. One of the key challenges in AOI is optimising inspection tolerances. Traditionally, this process relies heavily on the expertise and intuition of engineers, making it subjective and prone to inconsistency. To address this, we are developing an intelligent, data-driven approach to optimise inspection tolerances in a more objective and consistent manner. Most existing research in this area focuses primarily on minimising false calls, often at the risk of allowing actual defects to go undetected. This oversight can compromise product quality, especially in critical sectors such as medical, defence, and automotive industries. Our approach introduces the use of percentile rank, amongst other logical strategies, to ensure that genuine defects are not overlooked. With continued refinement, our method aims to reach a point where every flagged item is a true defect, thereby eliminating the need for manual inspection. Our proof of concept achieved an 18% reduction in false calls at the 80th percentile rank, while maintaining a 100% recall rate. This makes the system both efficient and reliable, offering significant time and cost savings.

eess.SP↗

arXiv2023

Learning to Predict Grip Quality from Simulation: Establishing a Digital Twin to Generate Simulated Data for a Grip Stability Metric

A robust grip is key to successful manipulation and joining of work pieces involved in any industrial assembly process. Stability of a grip depends on geometric and physical properties of the object as well as the gripper itself. Current state-of-the-art algorithms can usually predict if a grip would fail. However, they are not able to predict the force at which the gripped object starts to slip, which is critical as the object might be subjected to external forces, e.g. when joining it with another object. This research project aims to develop a AI-based approach for a grip metric based on tactile sensor data capturing the physical interactions between gripper and object. Thus, the maximum force that can be applied to the object before it begins to slip should be predicted before manipulating the object. The RGB image of the contact surface between the object and gripper jaws obtained from GelSight tactile sensors during the initial phase of the grip should serve as a training input for the grip metric. To generate such a data set, a pull experiment is designed using a UR 5 robot. Performing these experiments in real life to populate the data set is time consuming since different object classes, geometries, material properties and surface textures need to be considered to enhance the robustness of the prediction algorithm. Hence, a simulation model of the experimental setup has been developed to both speed up and automate the data generation process. In this paper, the design of this digital twin and the accuracy of the synthetic data are presented. State-of-the-art image comparison algorithms show that the simulated RGB images of the contact surface match the experimental data. In addition, the maximum pull forces can be reproduced for different object classes and grip scenarios. As a result, the synthetically generated data can be further used to train the neural grip metric network.

cs.RO↗

arXiv2022

Epileptic Seizure Classification Using Combined Labels and a Genetic Algorithm

Epilepsy affects 50 million people worldwide and is one of the most common serious neurological disorders. Seizure detection and classification is a valuable tool for diagnosing and maintaining the condition. An automated classification algorithm will allow for accurate diagnosis. Utilising the Temple University Hospital (TUH) Seizure Corpus, six seizure types are compared; absence, complex partial, myoclonic, simple partial, tonic and tonic- clonic models. This study proposes a method that utilises unique features with a novel parallel classifier - Parallel Genetic Naive Bayes (NB) Seizure Classifier (PGNBSC). The PGNBSC algorithm searches through the features and by reclassifying the data each time, the algorithm will create a matrix for optimum search criteria. Ictal states from the EEGs are segmented into 1.8 s windows, where the epochs are then further decomposed into 13 different features from the first intrinsic mode function (IMF). The features are compared using an original NB classifier in the first model. This is improved upon in a second model by using a genetic algorithm (Binary Grey Wolf Optimisation, Option 1) with a NB classifier. The third model uses a combination of the simple partial and complex partial seizures to provide the highest classification accuracy for each of the six seizures amongst the three models (20%, 53%, and 85% for first, second, and third model, respectively).

eess.SP↗

arXiv2021

Seizure Classification of EEG based on Wavelet Signal Denoising Using a Novel Channel Selection Algorithm

Epilepsy is a disorder of the nervous system that can affect people of any age group. With roughly 50 million people worldwide diagnosed with the disorder, it is one of the most common neurological disorders. The EEG is an indispensable tool for diagnosis of epileptic seizures in an ideal case, as brain waves from an epileptic person will present distinct abnormalities. However, in real world situations there will often be biological and electrical noise interference, as well as the issue of a multichannel signal, which introduce a great challenge for seizure detection. For this study, the Temple University Hospital (TUH) EEG Seizure Corpus dataset was used. This paper proposes a novel channel selection method which isolates different frequency ranges within five channels. This is based upon the frequencies at which normal brain waveforms exhibit. A one second window was selected, with a 0.5 second overlap. Wavelet signal denoising was performed using Daubechies 4 wavelet decomposition, thresholding was applied using minimax soft thresholding criteria. Filter banking was used to localise frequency ranges from five specific channels. Statistical features were then derived from the outputs. After performing bagged tree classification using 500 learners, a test accuracy of 0.82 was achieved.

q-bio.PE↗

arXiv2021

Assessing the effects of time-dependent restrictions and control actions to flatten the curve of COVID-19 in Kazakhstan

This paper presents the assessment of time-dependent national-level restrictions and control actions and their effects in fighting the COVID-19 pandemic. By analysing the transmission dynamics during the first wave of COVID-19 in the country, the effectiveness of the various levels of control actions taken to flatten the curve can be better quantified and understood. This in turn can help the relevant authorities to better plan for and control the subsequent waves of the pandemic. To achieve this, a deterministic population model for the pandemic is firstly developed to take into consideration the time-dependent characteristics of the model parameters, especially on the ever-evolving value of the reproduction number, which is one of the critical measures used to describe the transmission dynamics of this pandemic. The reproduction number alongside other key parameters of the model can then be estimated by fitting the model to real-world data using numerical optimisation techniques or by inducing ad-hoc control actions as recorded in the news platforms. In this paper, the model is verified using a case study based on the data from the first wave of COVID-19 in the Republic of Kazakhstan. The model is fitted to provide estimates for two settings in simulations; time-invariant and time-varying (with bounded constraints) parameters. Finally, some forecasts are made using four scenarios with time-dependent control measures so as to determine which would reflect on the actual situations better.

q-bio.PE↗

arXiv2020

COVID-19: Development of a Robust Mathematical Model and Simulation Package with Consideration for Ageing Population and Time Delay for Control Action and Resusceptibility

The current global health emergency triggered by the pandemic COVID-19 is one of the greatest challenges mankind face in this generation. Computational simulations have played an important role to predict the development of the current pandemic. Such simulations enable early indications on the future projections of the pandemic and is useful to estimate the efficiency of control action in the battle against the SARS-CoV-2 virus. The SEIR model is a well-known method used in computational simulations of infectious viral diseases and it has been widely used to model other epidemics such as Ebola, SARS, MERS, and influenza A. This paper presents a modified SEIRS model with additional exit conditions in the form of death rates and resusceptibility, where we can tune the exit conditions in the model to extend prediction on the current projections of the pandemic into three possible outcomes; death, recovery, and recovery with a possibility of resusceptibility. The model also considers specific information such as ageing factor of the population, time delay on the development of the pandemic due to control action measures, as well as resusceptibility with temporal immune response. Owing to huge variations in clinical symptoms exhibited by COVID-19, the proposed model aims to reflect better on the current scenario and case data reported, such that the spread of the disease and the efficiency of the control action taken can be better understood. The model is verified using two case studies for verification and prediction studies, based on the real-world data in South Korea and Northern Ireland, respectively.

q-bio.PE↗