Papers - THI THI ZIN
-
Motion Detection Method for Reducing Foreground Aperture Problem in Background Modelling Reviewed
Thi Thi Zin, Pyke Tin, Cho Nilar Phyo
LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies 260 - 261 2020.3
Authorship：Lead author, Corresponding author Language：English Publishing type：Research paper (international conference proceedings) Publisher：LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies
Motion detection plays as an important role in the implementation of video surveillance system and video analysis applications. The detection of moving foreground objects from the complex background scene is the first step in most of the computer vision based surveillance applications. In this paper, we present a new motion detection method using background modelling technique and moving average feature. For establishing the robust background model, we create the Gamma Mixture Model based on the Gamma distribution function. For handling the foreground aperture problem, we use the feature called moving average which can well recognize the alive silent objects. As post-processing, we handle the shadow removing process by generating the dynamic shadow threshold. The experimental results show that the proposed motion detection method can detect the accurate foreground object even though the foreground object appears as silent background objects in real-time environment.
-
Handwritten Characters Segmentation using Projection Approach Reviewed
Thi Thi Zin, Shin Thant, Ye Htet, Pyke Tin
LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies 107 - 108 2020.3
Authorship：Lead author, Corresponding author Language：English Publishing type：Research paper (international conference proceedings) Publisher：LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies
In the area of optical character recognition, handwritten character segmentation is still an ongoing process. Having good segmentation result can provide the better recognition accuracy. In the proposed system, segmentation is carried out mainly on labelling and projection concepts. The input word is firstly labelled. Then, the modified word is segmented with projection approach. The experiments are performed on local dataset with 1600 words approximately and the system gets segmentation accuracy around 85.75 percentage.
-
Cow Identification System using Ear Tag Recognition Reviewed
Thi Thi Zin, S. Misawa, Moe Zet Pwint, Shin Thant, Pann Thinzar Seint, K. Sumi, K. Yoshida
LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies 65 - 66 2020.3
Authorship：Lead author, Corresponding author Language：English Publishing type：Research paper (international conference proceedings) Publisher：LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies
In precision dairy farming, the valid record of individual cow identification is an important factor in large herds management. In this paper, we propose a cow's ear tag recognition system that can be used in dairy cow management. Firstly, cow's head detection is performed by using You Only Look Once (YOLO) object detector followed by ear tag recognition. The ear tag extraction and recognition processes are carried out by image processing techniques and Convolutional Neural Network (CNN) classifier on detected cow's head images. The experiments are conducted by using videos from dairy farm at Hokkaido prefecture, Japan. The proposed system achieved the reliable results which will support to give the informative status in smart farming.
-
Human Action Analysis Using Virtual Grounding Point and Motion History Reviewed
Swe Nwe Nwe Htun, Thi Thi Zin, H. Hama
LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies 249 - 250 2020.3
Language：English Publishing type：Research paper (international conference proceedings) Publisher：LifeTech 2020 - 2020 IEEE 2nd Global Conference on Life Sciences and Technologies
In this paper, we propose an approach to human action analysis for home care monitoring system in the aspect of image processing and life science technologies. We introduce a new concept of a virtual grounding point representing the position of a target person as an innovated feature for action analysis. Specifically, in developing action analysis, the background subtraction is firstly conducted by applying the Mixture of Gaussian and low rank subspace learning. After that, the graph cut is embedded to enhance the foregrounds in order to detect both of moving and motionless object. Secondly, the virtual grounding point is calculated by using the centroid of silhouette image. Finally, motion of the person is estimated by using timed motion history image in order to improve the accuracy of action analysis. A series of the experiments are conducted to confirm the effectiveness of the proposed method.
-
Some Aspects of Mathematical Modeling Techniques in Dairy Science Reviewed
Thi Thi Zin, Pyke Tin and H. Hama
International Workshop on Frontiers of Computer Vision (IW-FCV) 1 - 8 2020.2
Language：English Publishing type：Research paper (international conference proceedings)
Today world is floating in the ocean of data science which is a multi-disciplinary approach to every aspect of our better life. Agriculture 4.0 is coming in and Industry 4.0 is moving forward so that second Information and Communication revolutionary becomes to play a key role in this era. On the other hands, consumers are demanding more quality dairy products than before. Dairy farmers are eager to use the information and communication technologies that push the precision dairy farming to frontier innovations. Due to a shortage of labor in dairy farms and at the same time the farm sizes are bigger, the utilization of ICT along with artificial intelligence and the internet of things are becoming appropriate technologies. Whatever it may be, fundamentally we need to do some preliminary research prior to large scale and wide applications to the real world. In almost analytical and empirical research, we are trying in the ocean of data science. These data could not be used efficiently unless we do further processing and projections. The raw data need to be processed to obtain useful information and knowledge, which could be used for important on-farms decisions. This kind of process is called mathematical modeling which could support to decision-makers to produce correct and valid decisions. In this paper, we will present and analyze some mathematical models in the frame works of dairy science. Moreover, we shall illustrate how these models could be used in practical ways.
-
Background Modelling Using Temporal Average Filter and Running Gaussian Average Reviewed
Thi Thi Zin, Cho Cho Mar and K. Sumi
International Workshop on Frontiers of Computer Vision (IW-FCV) 1 - 8 2020.2
Language：English Publishing type：Research paper (international conference proceedings)
Object detection is very important and fundamental stage for further post processing stages such as individual identification, action recognition, tracking and behavior analysis. Traditional object detection methods start from back-ground modelling stage. Background modelling is a complex and challenging task which highly depends on the speed of moving objects (foreground objects) and stability of static regions (background region). In this paper, we proposed pixel level background modelling method. In this background modelling task, the two main components of background modelling are proposed; (1) background model initialization by selecting the dynamic frames from video sequence and building the model with Temporal Average filter and (2) updating the back-ground model by using Running Gaussian Average method. These methods are applied to RGB images.
-
The Body Condition Score Indicators for Dairy Cows Using 3D Camera Reviewed
Thi Thi Zin, Pann Thinzar Seint, Pyke Tin, Y. Hori
International Workshop on Frontiers of Computer Vision (IW-FCV) 1 - 8 2020.2
Language：English Publishing type：Research paper (international conference proceedings)
Body Condition Score (BCS) evaluates energy reserves of cows when making nutritional management. This information is also useful in fertility rate and milk production. Keeping cows in appropriate condition throughout the production cycle can improve reproductive efficiency and positive impact for economics. In this paper, we propose the effective indicators for automated BCS measurement system and the experimental data are collected at the milking station. We compute the variation of angular area from automatic anatomical point extraction and learning parameters on Region of Interest (ROI) from 3D information. The statistical analysis of BCS trend on year-round data are presented to give the detailed illustrations for dairy farmers. According to the recorded BCS trends, the proposed method can validate that the lower BCS were obtained in early stage of lactation period rather than other seasons.
-
A Hybrid Visual Stochastic Approach to Dairy Cow Monitoring System Reviewed
Thi Thi Zin, Pyke Tin, I. Kobayashi, H Hama
Transactions on Engineering Technologies 119 - 129 2020.1
Language：English Publishing type：Research paper (scientific journal)
In the era of fourth industrial revolution or Industry 4.0 together with the second information and communication technology era, the way we human beings live, act and do are always challenging by the waves of new technologies such as Internet of Things, Artificial Intelligence, Cloud Computing so on and so forth. These new technologies also drive life science, academic, business and production industries for better sides. Among them, one of challenging and innovative technology in life science industry is precision dairy farming which has been pushed into frontline topic among the academia and industry farm managers. Generally speaking, the precision dairy farming can be defined as the use of technologies advanced or simple to analyze the physical and mental behaviors of an individual cows for specifying health and profitability indicators so that overall management and farm performance are to be improved. In other words, the precision dairy farming will focus on welfare and health care for making the farm returns optimal through the use of technologies. Here, technologies may range from daily milk recording to automatic body conditions scoring for individual cows and an accurate prediction of calving time and reporting unusual occurrences during the delivery times.
In order to realize the objectives of precision dairy farming as a fundamental step, a hybrid visual stochastic approach which is a fusion of image technology and statistical method to dairy cow monitoring system is introduced in this paper. The proposed system will investigate four key areas for the precision dairy farming namely: cow identification, body condition scoring, detection of estrus behavior, prediction of time for the occurrence calving event. In doing so, the combination of image technology, statistical methods and stochastic models will be utilized. Specifically, image processing methods will be performed to detect cow activities such as standing, lying and walking in association with time, space and frequencies. Then collected data are to be transformed into a stochastic model of special type of Markov Chain for decision making process. Then some experimental results will be shown by using both image and statistical data collected from the real life environments and some available datasets. -
Robust tracking of cattle using super pixels and local graph cut for monitoring systems Reviewed
Yukie Hashimoto, Hiromitsu Hama, ThiThi Zin
International Journal of Innovative Computing, Information and Control ６（４） 1469 - 1475 2020
Authorship：Last author Language：English Publishing type：Research paper (scientific journal)
-
A Correlated Random Walk Modeling Method for Dairy Cow Inter-calving Body Condition Score Pattern Analysis Reviewed
ThiThi Zin, Pyke Tin
ICIC Express Letters, 14 2020
Authorship：Lead author, Corresponding author Language：English Publishing type：Research paper (scientific journal)
-
Robust Tracking of Cattle Using Super Pixels and Local Graph Cut for Monitoring Systems Reviewed
ThiThi Zin
International Journal of Innovative Computing, Information and Control (IJICIC) 16 2020
Publishing type：Research paper (scientific journal)
-
Motion History and Shape Orientation Based Human Action Analysis Reviewed
Swe Nwe Nwe Htun, Thi Thi Zin
2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019 754 - 755 2019.10
Authorship：Corresponding author Language：English Publishing type：Research paper (international conference proceedings) Publisher：2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019
In recent decades, many research works focused on the considerations of falling and post-falling event analysis in the aspect of image processing technology to meet the consumer perspective of users. In this paper, the more informative considerations of human action analysis are developed for the prior state of fall or normal actions. In doing so, the effective background subtraction namely the Mixture of Gaussian with Low-Rank Matrix Factorization model is used to obtain robust and adaptive foreground from the practical video sequences. Then, the spatial-temporal bilateral grid for video sequences is constructed by using a standard graph cut theory to improve the cleaned foreground in order to detect the moving/motionless object region. Then, human actions are analyzed by employing motion history and shape orientation using the approximated ellipse method. The experiments are conducted on publicly available video sequences and our simulated video sequences.
-
Image-Based Feeding Behavior Detection for Dairy Cow Reviewed
K. Shiiya, F. Otsuka, Thi Thi Zin, I. Kobayashi
2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019 756 - 757 2019.10
Language：English Publishing type：Research paper (international conference proceedings) Publisher：2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019
Feeding behavior is an important source of information to know cow's health status, because it is influenced by the feeding environment, cow's physiological changes and health conditions. However, a cow feeds intermittently throughout the day, it is difficult to measure feeding time by visual measurement at field level. In this paper we propose a measurement method of feeding frequency and feeding time for dairy cow, by detecting the feeding behavior by non-contact using a camera.
-
Handwritten Character Segmentation in Tablet Based Application Reviewed
Myat Thiri Wai, Thi Thi Zin, M. Yokota, Khin Than Mya
2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019 760 - 761 2019.10
Language：English Publishing type：Research paper (international conference proceedings) Publisher：2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019
Nowadays, modern tablets are widely used in the contribution of universal access to education, equality in the exercise of teaching and learning, aiming for more efficient management and administration. The recognition of handwritten characters on these tablets become a necessary consequence in this technology age. The main challenging of handwritten character recognition is splitting each character of unrestricted handwritten scripts and there is no complete solution yet according to our knowledge. In this system, the combination of sliding windows, Region of Interest (ROI) box and Convolutional Neural Network (CNN) are used to execute implicit segmentation of handwritten characters. For performing the experiments, our own dataset is constructed by collecting handwritten data from 24 members of our laboratory using three different tablet PC models.
-
Feature Analysis for Non Formal Education Project in Myanmar Reviewed
Mie Mie Khin, Mie MIe Tin, Thi Thi Zin, Pyke Tin
2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019 890 - 891 2019.10
Language：English Publishing type：Research paper (international conference proceedings) Publisher：2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019
In a digitally mediated world, Literacy is vital to live. IT based Digital technology is the creation of quality education. It is the main pillar to develop a sustainable socioeconomic development in the country. Myanmar also starts to develop E-government system for every sector. This paper describes feature selection for non-formal education system using literacy project. With this approach, we consider that feature selection of linear regression the analysis is more suitable than support vector method. We are also continuing analysis of this e-education system for rural area in Mandalay division of Myanmar. This paper analysis based on big data concept.
-
Emotion Analysis of Twitter Users on Natural Disasters Reviewed International coauthorship
Nann Hwan Khun, Thi Thi Zin M. Yokota Hnin Aye Thant
2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019 342 - 343 2019.10
Authorship：Corresponding author Language：English Publishing type：Research paper (international conference proceedings) Publisher：2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019
In this information era, people usually express their views and emotions on a wide range of topics through social networking sites and so the role of emotion analysis in social media has been the subject of considerable research. The idea behind this research is that the emotions people express in their status updates can tell us something about how their emotions fluctuate day-to-day due to natural disasters. In this paper, we targeted for emotion analysis of Twitter users on natural disasters. By identifying these emotions, we can help first responders for better managing the situations such as mental health of survived victims. Our experiment is based on California Camp Fire that is happened in 2018 November.
-
Consumer Technology Perspective of a Trinomial Random Walk Model Reviewed
Thi Thi Zin, Pyke Tin, H. Hama
2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019 758 - 759 2019.10
Authorship：Lead author, Corresponding author Language：English Publishing type：Research paper (international conference proceedings) Publisher：2019 IEEE 8th Global Conference on Consumer Electronics, GCCE 2019
Recently, the consumer technology association has widened the horizons of the consumer electronic society to include a variety of technologies such as the Internet of Things, Artificial Intelligence, vertical farming and etc. Moreover, among many others, the consumer technology highlights the importance of organic dairy cows and they are well beings so that timely based energy balance and body condition scores of individual cows are needed to be thoroughly studied. In this paper, we shall introduce a Trinomial Random Walk Model as an illustration of consumer technology for analyzing changes in dairy cow body conditions from time to time. In doing so, we will use the 5-point scale system with increment 0.25. Some experimental simulation results are presented by varying the parameters involved.
-
Image Technology and Matrix-Geometric Method for Automatic Dairy Cow Body Condition Scoring Reviewed
Thi Thi Zin, Pyke Tin, Y. Horii, I. Kobayashi
International Journal of Biomedical Soft Computing and Human Sciences 24 ( 1 ) 2019.7
Language：English Publishing type：Research paper (scientific journal)
Dairy cow body condition scoring is a key and important in nearly every aspect of modern precision dairy farms for management of welfare and healthcare, milk yields and split meals (feeding system), aware heat and reproduction peak, calving in smooth and calf saving in proof so that all in all profitability in use. However, the majority of dairy farmers are not doing the body condition scoring on regular basis due to lack of automation so that it causes time-consuming and very subjective. Thus in this paper, an automatic and efficient dairy cow body scoring system is proposed by using Image Technology and Matrix-Geometric Method. By fusing image technol-ogy and matrix-geometric method as a hybrid can lead to a new and efficient technique of dairy cow body condition scoring system. In order to do so, firstly, the anatomical cow body points are extracted from the back views and top views of cow images. Then the geometric properties of the extracted anatomical points are transformed into a Markov Chain Matrix to determine the body condition scores. For confirmation of the validity of the proposed method, some experimental results are shown by using a public cow body condition database.
-
Complex Human–Object Interactions Analyzer Using a DCNN and SVM Hybrid Approach Reviewed
Cho Nilar Phyo, Thi Thi Zin and Pyke Tin
Applied Sciences (Switzerland) 9 ( 9 ) 2019.5
Language：English Publishing type：Research paper (scientific journal) Publisher：Applied Sciences (Switzerland)
Nowadays, with the emergence of sophisticated electronic devices, human daily activities are becoming more and more complex. On the other hand, research has begun on the use of reliable, cost-effective sensors, patient monitoring systems, and other systems that make daily life more comfortable for the elderly. Moreover, in the field of computer vision, human action recognition (HAR) has drawn much attention as a subject of research because of its potential for numerous cost-effective applications. Although much research has investigated the use of HAR, most has dealt with simple basic actions in a simplified environment; not much work has been done in more complex, real-world environments. Therefore, a need exists for a system that can recognize complex daily activities in a variety of realistic environments. In this paper, we propose a system for recognizing such activities, in which humans interact with various objects, taking into consideration object-oriented activity information, the use of deep convolutional neural networks, and a multi-class support vector machine (multi-class SVM). The experiments are performed on a publicly available cornell activity dataset: CAD-120 which is a dataset of human-object interactions featuring ten high-level daily activities. The outcome results show that the proposed system achieves an accuracy of 93.33%, which is higher than other state-of-the-art methods, and has great potential for applications recognizing complex daily activities.
DOI： 10.3390/app9091869
-
Deep Learning for Recognizing Human Activities Using Motions of Skeletal Joints Reviewed
Cho Nilar Phyo, Thi Thi Zin, Pyke Tin
IEEE Transactions on Consumer Electronics 65 ( 2 ) 243 - 252 2019.5
Language：English Publishing type：Research paper (scientific journal) Publisher：IEEE Transactions on Consumer Electronics
With advances in consumer electronics, demands have increased for greater granularity in differentiating and analyzing human daily activities. Moreover, the potential of machine learning, and especially deep learning, has become apparent as research proceeds in applications, such as monitoring the elderly, and surveillance for detection of suspicious people and objects left in public places. Although some techniques have been developed for human action recognition (HAR) using wearable sensors, these devices can place unnecessary mental and physical discomfort on people, especially children and the elderly. Therefore, research has focused on image-based HAR, placing it on the front line of developments in consumer electronics. This paper proposes an intelligent HAR system which can automatically recognize the human daily activities from depth sensors using human skeleton information, combining the techniques of image processing and deep learning. Moreover, due to low computational cost and high accuracy outcomes, an approach using skeleton information has proven very promising, and can be utilized without any restrictions on environments or domain structures. Therefore, this paper discusses the development of an effective skeleton information-based HAR which can be used as an embedded system. The experiments are performed using two famous public datasets of human daily activities. According to the experimental results, the proposed system outperforms other state-of-the-art methods on both datasets.