CHChin-Kuan Ho

Chin-Kuan Ho

*This is an algorithmically generated author record based on publicly available information.

Chin-Kuan Ho has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Convxpf-Net: An Explainable Ai Approach For Recognizing Pedestrian Behaviour Using Convolutional Xgboost Pyramid Fractional Network

    by Amol Raoji Madane, Ho Chin Kuan

    In Advanced Driver Assistance System (ADAS), identifying pedestrians and predicting their future behaviour is essential for preventing accidents, however during heavy occlusion, the prevailing Pedestrian Behaviour Detection (PBD) techniques fails to identify the appropriate behaviour. Therefore, in this research, a novel Explainable AI based Convolutional XGBoost pyramid fractional Network (ConvXPF-Net) is proposed. The existing PBD techniques struggles to identify the patterns in the frames due to orientation-invariant regression head, and hence, YOLOv8 based Visual Casual Model (Y8-VCM) is employed, which eliminates the Boundary Discontinuity problem in Angle prediction. Besides, for addressing missed detections and issues with distracted walking during classification, a Convolutional XGBoost Pyramid Fractional Network (ConvXPF-Net) is utilized. The complexities in detecting the pedestrians with less radial velocities due to minimal Doppler shift, is reduced by Pyramid Net Technique (PNT), and then for recognizing the subtle behavioural changes that leads to oversimplification of the features, a Fractional Calculus Technique is employed. Further, based on the classified PB, the movement of pedestrians is determined, which further aids in performing speed and brake automation. Finally, the proposed work is validated using Explainable AI, and the outcomes attained shows the effectiveness of proposed work in Pedestrian behaviour prediction.

  • Front-end deep learning web apps development and deployment: a review

    by Hock-Ann Goh, Chin-Kuan Ho, Fazly Salleh Abas

    Machine learning and deep learning models are commonly developed using programming languages such as Python, C++, or R and deployed as web apps delivered from a back-end server or as mobile apps installed from an app store. However, recently front-end technologies and JavaScript libraries, such as TensorFlow.js, have been introduced to make machine learning more accessible to researchers and end-users. Using JavaScript, TensorFlow.js can define, train, and run new or existing, pre-trained machine learning models entirely in the browser from the client-side, which improves the user experience through interaction while preserving privacy. Deep learning models deployed on front-end browsers must be small, have fast inference, and ideally be interactive in real-time. Therefore, the emphasis on development and deployment is different. This paper aims to review the development and deployment of these deep-learning web apps to raise awareness of the recent advancements and encourage more researchers to take advantage of this technology for their own work. First, the rationale behind the deployment stack (front-end, JavaScript, and TensorFlow.js) is discussed. Then, the development approach for obtaining deep learning models that are optimized and suitable for front-end deployment is then described. The article also provides current web applications divided into seven categories to show deep learning potential on the front end. These include web apps for deep learning playground, pose detection and gesture tracking, music and art creation, expression detection and facial recognition, video segmentation, image and signal analysis, healthcare diagnosis, recognition, and identification.


  • Open Access

    WSFeIn: A Novel, Dynamic Web Service Composition Adapter for Cloud-Based Mobile Application

    by R. Kanesaraj Ramasamy, Fang-Fang Chua, Su-Cheng Haw, Chin-Kuan Ho

    Service-Oriented Computing (SOC) has been a cornerstone development in today’s fast-paced world, covering the lifecycle of services and contributing to service delivery through distributed applications. SOC is striving to integrate cloud computing with complicated mobile apps. Dynamic and adaptable web service composition is the tip of the iceberg for SOA adoption. Dynamic service binding is vital for mobile computing due to the need for distributed mobile internet consumption at runtime. This study addresses SOC difficulties associated with web service composition, whose growth creates a paradigm change in identifying data type matching solutions. This allows for data type-level matching research to ensure high-quality web service creation. In this work, the composition process is divided into three phases: web service discovery, web service selection, and web service composition, where web service personalization and workflow reliability are emphasized. The final result is a complicated mobile app that runs without depleting device resources and an adaptable, reusable web service composition workflow. This improves matching at the data type level, an SOC pain point.