Papers by Nikom Suvonvorn
This paper deals with designing algorithms based on the "stable marriages" paradigm, for them to ... more This paper deals with designing algorithms based on the "stable marriages" paradigm, for them to take additional constraints into account. First algorithms scan the so-called "marriage table" to optimize "global satisfaction" and "equity" over all couples. Then, we introduce an hybrid version between the Gale-Shappley classical algorithm and ours. Results on thousands of populations, up to 200 person large, are systematically evaluated. These algorithms progressively improve both satisfaction and equity but they do not guarantee complete stability. Thus, the BZ algorithm is made to blow blockages out. It still produces about 5% instable results in average. After a case study that sorts blocking situations into 4 types, we explain how to resolve instability in forcing blocking pairs to marry wrt. their type. The resulting S-procedure applies after every previous algorithm, and results are systematically compared to Gale-Shapley's on 1000 instances of a 200-person-large population. In conclusion, a more experimental study of the complexity allows to suggest trade off criterions between maximal quality and rapidity.
International Journal of Computer and Electrical Engineering, 2012
Our previous work proposed hand posture estimation technique. The hand region is first extracted ... more Our previous work proposed hand posture estimation technique. The hand region is first extracted using depth image, and then the initial features, such as fingertip, hand center point, and palm size, have been calculated. The concept of active contour using energy function is implemented in order to track fingertip position in the frame image sequence. To discriminate the hand posture sets, a hand feature definition have been established, which is composed of finger shape and hand appearance features. The features are defined as characteristic code to represent hand posture. This technique is applied to the American finger-spelling. Since we focused on Thai letters fingerspelling which are based on American finger-spelling hand postures. Therefore, this paper has introduced a hand posture estimation method for Thai letters finger-spelling. The Hidden Markov Model (HMM) method is used to build a learning model to recognize the sequence of the American finger-spelling hand postures and to provide the Thai letter finger-spelling. The performance of the recognition system can be measured at around 70% recognition rate.
In this paper, we introduce a method for fingerspelling recognition system. The objective is to h... more In this paper, we introduce a method for fingerspelling recognition system. The objective is to help the deaf or non-vocal persons to improve their skills on the finger-spelling. Many researches in this field have proposed methods mostly based on hand posture estimation techniques. We propose an alternative flexible method based on fuzzy finger shape and hand appearance analysis. By using depth image, the hand is extracted and tracked using an active contour like method. Its features, such as, finger shape, and hand appearance, have been defined as chain code, which are input to the American finger-spelling recognition system by using a vote method. The performance of the system is tested in real-time environment, which results in around 70% recognition rate.
Springer eBooks, Dec 15, 2013
This paper presents a real time estimation method for 3D trajectory of fingertips. Our approach i... more This paper presents a real time estimation method for 3D trajectory of fingertips. Our approach is based on depth vision, with Kinect depth sensor. The hand is extracted using hand detector and depth image from sensor. The fingertips are located by the analysis of the curvature of hand contour. The fingertips detector is implemented using concept of active contour which combine the energy of continuity, curvature, direction, depth and distance. The trajectory of fingertips is filtered to reduce the tracking error. The experiment is evaluated on the fingers movement sequences. Besides, the capabilities of the method are demonstrated on the real-time Human-Computer Interaction (HCI) application.
Top-View based Human Action Recognition using Depth and Color Information
International Journal of Power Electronics and Drive Systems, Jun 1, 2023
The investigation of a deep neural network for pedestrian classification using transfer learning ... more The investigation of a deep neural network for pedestrian classification using transfer learning methods is proposed in this study. The development of deep convolutional neural networks has significantly improved the autonomous driver assistance system for pedestrian classification. However, the presence of partially occluded parts and the appearance variation under complex scenes are still robust to challenge in the pedestrian detection system. To address this problem, we proposed six transfer learning models: end-to-end convolutional neural network (CNN) model, scratch-trained residual network (ResNet50) model, and four transfer learning models: visual geometry group 16 (VGG16), GoogLeNet (InceptionV3), ResNet50, and MobileNet. The performance of the pedestrian classification was evaluated using four publicly datasets:
Lecture Notes in Computer Science, 2011
Traditional motion-capture systems have found widespread use in content creation in the entertain... more Traditional motion-capture systems have found widespread use in content creation in the entertainment industry. However, their use as real-time user-input devices has been limited by their expense. In this thesis work, we propose an easy-to-use and inexpensive system that facilitates 3-D articulated user-input using the hands, enabling new user-interfaces and applications. Our system optically tracks an ordinary cloth glove that is imprinted with a custom pattern. Observing a custom pattern simplifies the pose estimation problem, allowing us to eschew computationally demanding inference algorithms in favor of a simple and fast nearest-neighbor approach. We investigate optimal design of the cloth pattern as well as possible applications for our system.
Contextual Data Modeling for Recommender System in Building and Construction Materials Business
2022 26th International Computer Science and Engineering Conference (ICSEC)
This paper deals with designing algorithms based on the "stable marriages" paradigm, for them to ... more This paper deals with designing algorithms based on the "stable marriages" paradigm, for them to take additional constraints into account. First algorithms scan the so-called "marriage table" to optimize "global satisfaction" and "equity" over all couples. Then, we introduce an hybrid version between the Gale-Shappley classical algorithm and ours. Results on thousands of populations, up to 200 person large, are systematically evaluated. These algorithms progressively improve both satisfaction and equity but they do not guarantee complete stability. Thus, the BZ algorithm is made to blow blockages out. It still produces about 5% instable results in average. After a case study that sorts blocking situations into 4 types, we explain how to resolve instability in forcing blocking pairs to marry wrt. their type. The resulting S-procedure applies after every previous algorithm, and results are systematically compared to Gale-Shapley's on 1000 instances of a 200-person-large population. In conclusion, a more experimental study of the complexity allows to suggest trade off criterions between maximal quality and rapidity.
this research proposed the human activityrecognition system that described complete flow of proce... more this research proposed the human activityrecognition system that described complete flow of processes fromlowest process (dealing with images) to highest process (recognizehuman activity). We proposed human action recognition thatmanage image sequence then recognize human action with simplehuman model by model-based recognition technique. Theexperimental result shows good accuracy which up to 93%correctly recognized. We proposed the human activity processwith 3 methods that consecutive improved. All of those methodscan use the result of action recognition as inputs. First method isFSM recognizer. The human model in Finite State Machine (FSM)recognizer can be modeled by rational condition that make it easyto understand and consume low computation cost but it hard todefine complex activity condition so it is unsuitable method forcomplex activity. The second recognizer applied Hidden MarkovModel (HMM) for activity modeling. The HMM recognizer candealing with much more complex activity ...
American Astronomical Society Meeting Abstracts #235, 2020
The research aimed to identify the extent of practicing creative thinking skills among students o... more The research aimed to identify the extent of practicing creative thinking skills among students of educational technology at the University of Ibb, Yemen in applying innovative educational technologies. To achieve the objectives of the study and answer its questions, the researchers used the descriptive method. The study population was educational technology students at the University of Ibb. The research sample involved (50) students of the second level in the Department of Educational Technology in the academic year 2018/ 2019. A scale of creative thinking and an open-ended questionnaire were used to assess creative thinking skills and identify the obstacles of creative thinking among students. The results showed a clear decline in the students' ability to practice creative thinking skills with a general average of (34.90) for group A students, and an average of (35.15) for group B students. The study revealed a number of obstacles of developing students' creative thinking. The study concluded with a set of recommendations and suggestions.
ECTI Transactions on Computer and Information Technology (ECTI-CIT), 2021
Fall detection techniques for helping the elderly were developed based on identifying falling sta... more Fall detection techniques for helping the elderly were developed based on identifying falling states using simulated falls. However, some real-life falling states were left undetected, which led to this work on analysing falling states. The aim was to find the differences between active daily living and soft falls where falling states were undetected. This is the first consideration to be based on the threshold-based algorithms using the acceleration data stored in an activity database. This study addresses soft falls in addition to the general falls based on two falling states. Despite the number of false alarms being higher rising from 18.5% to 56.5%, the sensitivity was increased from 52% to 92.5% for general falls, and from 56% to 86% for soft falls. Our experimental results show the importance of state occurrence for soft fall detection, and will be used to build a learning model for soft fall detection.
Garnet Geotherms of the Nyrbinsky and Boptuobinskaya Pipes Nakyn Kimberlite Fields Yuakutia
Lecture Notes in Computer Science, 2011
Traditional motion-capture systems have found widespread use in content creation in the entertain... more Traditional motion-capture systems have found widespread use in content creation in the entertainment industry. However, their use as real-time user-input devices has been limited by their expense. In this thesis work, we propose an easy-to-use and inexpensive system that facilitates 3-D articulated user-input using the hands, enabling new user-interfaces and applications. Our system optically tracks an ordinary cloth glove that is imprinted with a custom pattern. Observing a custom pattern simplifies the pose estimation problem, allowing us to eschew computationally demanding inference algorithms in favor of a simple and fast nearest-neighbor approach. We investigate optimal design of the cloth pattern as well as possible applications for our system.
A novel representation and algorithms for (quasi) stable marriages
International Conference on Informatics in Control, Automation and Robotics, 2005
Abstract—an on-line video processing for surveillance system is a very challenging problem. The c... more Abstract—an on-line video processing for surveillance system is a very challenging problem. The computational complexity of video analysis algorithms and the massive amount of data to be analyzed must be considered under real-time constraints. Moreover it needs to satisfy different criteria of application domain, such as, scalability, re-configurability, and quality of service. In this paper we propose a flexible/efficient video analysis framework for surveillance system which is a component-based architecture. The video acquisition, re-configurable video analysis, and video storage are some of the basic components. The component execution and inter-components synchronization are designed for supporting the multi-cores and multi-processors architecture with multi-threading implementation on.NET Framework. Experimental results on real-time motion tracking are presented with discussion. I.
Model-based Human Action Recognition
Abstract—The identification of human basic actions plays an important role for recognizing human ... more Abstract—The identification of human basic actions plays an important role for recognizing human activities in complex scene. In this paper we propose an approach for automatic human action recognition. The parametric model of human is extracted from image sequences using motion/texture based human detection and tracking. Action features from its model are carefully defined into the action interaction representation and used for the recognizing process. Performance of proposed method is tested experimentally using datasets under indoor environments.
Conventional motion capture systems have used widespread especially in the entertainment industry... more Conventional motion capture systems have used widespread especially in the entertainment industry. However, their usage as real time user input devices has been limited by their price. In this paper, we propose an easy-to-use and inexpensive system that facilitates 2-D user-input using video-based markers analysis captured from a consumer grade webcam. Our system detects and tracks a specific pattern of color marker on cloth glove for hand pose analysis. Some features are observed allowing us to recognize the particular commands. We evaluated the performance of system under real time constraint in real environment with a practical application.
Automatic monitoring of the activities of daily living (ADL) for elderly and disabled people from... more Automatic monitoring of the activities of daily living (ADL) for elderly and disabled people from image sequences is an important issue for homecare or care centre. Falls amongst
Uploads
Papers by Nikom Suvonvorn