Presentations -
-
Hybrid Approach: Markov Monte Carlo and Naïve Bayes Theorem to Predict Dairy Cow Calving Time International conference
Thi Thi Zin, Swe Zar Maw, Pyke Tin, Y. Horii and H. Hama
16th International Conference on Innovative Computing, Information and Control (ICICIC2022) (Online) 2022.9.15 ICIC International
Event date： 2022.9.15 - 2022.9.16
Language：English Presentation type：Oral presentation (general)
Venue：Online
In this paper, we propose a hybrid approach combing Markov Monte Carlo simulation and the Bayesian method to predict the time to the occurrence of calving events in dairy cows. A total of 30 individual dairy cows were monitored continuously under 24-hour video surveillance prior to calving. The video was annotated for the behaviors of lying, the transition from lying to standing, standing, and the transition from standing to lying for each cow from 72 hours and 168 hours prior to calving. We then derive the corresponding probabilities for each behavior. By using these probabilities, the Markov Monte Carlo simulations are performed to generate the behavior patterns of each cow prior to calving. We then investigate three types of datasets: actual, simulated, and a mixture of the two by using a Naïve Bayes Classifier to perform the prediction process. The experimental results showed that the hybrid approach correctly classified the calving event cent by cent.
-
Cattle Lameness Detection by Using Image Processing Technology International conference
T. Onizuka, Thi Thi Zin, I. Kobayashi
16th International Conference on Innovative Computing, Information and Control (ICICIC2022) (Online) 2022.9.15 ICIC International
Event date： 2022.9.15 - 2022.9.16
Language：English Presentation type：Oral presentation (general)
Venue：Online
Lameness is one of the three major diseases of dairy cows, mastitis, reproductive disorders, and limb and hoof diseases. Symptoms of lameness include pain and discomfort in the legs when walking, limping legs, and abnormal movements such as rounding of the back. Among the three major diseases, limb and hoof diseases in dairy cattle are needed to be detected in times, otherwise as the symptoms worsen, diseases such as mastitis and reproductive disorders may occur, leading to a decrease in productivity. Therefore, it is important to detect and treat cows with symptoms of lameness as early as possible. In this paper, we propose a method for cattle lameness detection using an RGB camera and analysing the walking behaviour of cows. Our proposed method can classify cows as healthy or lame with high accuracy. Moreover, we have performed a series of real-life experiments to confirm the effectiveness of the classification results comparing with lameness diagnosis by experts.
-
Comparative Study on Color Spaces, Distance Measures and Pretrained Deep Neural Networks for Cow Recognition International conference
Cho Cho Mar, Thi Thi Zin, Pyke Tin, I. Kobayashi, Y. Horii and K. Honkawa5
16th International Conference on Innovative Computing, Information and Control (ICICIC2022) (Online) 2022.9.15 ICIC International
Event date： 2022.9.15 - 2022.9.16
Language：English Presentation type：Oral presentation (general)
Venue：Online
As the cattle monitoring system plays as a crucial role in livestock production and dairy farming to cover the deficiencies in labor for continuous monitoring for each cattle. Robust object detection and tracking system boosts the performance and reliability of autonomous monitoring systems. The recognition task is also important to build the powerful tracking system. The main objective of this paper is to perform the recognition task by analyzing and comparing different color spaces and distance measures to choose the optimal ones. We extracted color moment features and co-occurrence matrix (CM) features from various color spaces: RGB, YCbCr, XYZ, HSV, CIElab, and gray level. We compare those features using different distance measures based on the accuracy values. We also make experiments on different Deep Learning pretrained networks to extract CNN (Convolutional Neural Network) features, and we compared the accuracy values on the classification results with Support Vector Machine (SVM) method. These experiments are tested on the cow dataset that are created from the 2 hours continuous video.
-
A Conceptual Framework and Operational Simulation Model for Cattle Lameness Detection International conference
Thi Thi Zin, Pyke Tin
16th International Conference on Innovative Computing, Information and Control (ICICIC2022) (Online) 2022.9.15 ICIC International
Event date： 2022.9.15 - 2022.9.16
Language：English Presentation type：Oral presentation (general)
Venue：Online
In this paper, we introduce a conceptual framework and operational simulation model for cattle lameness detection by using image depth data taken while cattle walking on the pathway from the milking center to the resting area. The proposed conceptual framework maps essential factors that determine lameness scores based on body movement variability. Specifically, body movement variability is to be defined as the root mean square successive differences, various types of information entropies, and geometric measures of the collected sequence of depth data during a field of depth camera view. In addition, we develop an operational simulation model that links the Monte Carlo simulation along with some popular probabilistic distribution functions such as uniform, normal, Poisson, and Gamma distributions for lameness status analysis. The simulation results lead to a conjecture that detection performance and the characteristics of lame and non-lame cows have a large influence on body movement variability. This conjecture will be realized by using real-life data in future work.
-
Gama-Markov Branching Process Approach to the Novel Coronavirus (COVID-19) Pandemic International conference
Thi Thi Zin, Pyke Tin and H. Hama
The Fifth International Symposium on Information and Knowledge Management (ISIKM2022) (onsite (Guangzhou, China / Kumamoto, Japan) and online) 2022.3.26 ICIC International
Event date： 2022.3.25 - 2022.3.26
Language：English Presentation type：Oral presentation (general)
Venue：onsite (Guangzhou, China / Kumamoto, Japan) and online
The spreads of the Novel Coronavirus (COVID-19) infectious disease have been increased exponentially in almost all countries around the world. Due to fast in transmission, big in mass and transform in variety, the world health organization has recognized the COVID-19 as a global pandemic. Since then, variety of measures such as mask wearing campaigns, social distance movements, mass vaccination programs and many others were taken globally to reduce the effective reproduction number of an infection below unity. In this paper, we propose a special type of Gama-Markov branching process model to monitor the reproduction mean rate by using data on outbreak size and outbreak duration. We develop a novel method for computing the
probabilities of the rate of spreads, outbreak sizes and outbreak duration. Some simulations based on the daily data appearing on the Internet are performed. The results show that the proposed model has promising potentials for the real-life applications. -
Predicting Calving Time of Dairy Cows by Autoregressive Integrated Moving Average (ARIMA) Model and Double Exponential Smoothing (DES) Model International conference
Tunn Cho Lwin, Thi Thi Zin and Pyke Tin
The Fifth International Symposium on Information and Knowledge Management (ISIKM2022) (onsite (Guangzhou, China / Kumamoto, Japan) and online) 2022.3.26 ICIC International
Event date： 2022.3.25 - 2022.3.26
Language：English Presentation type：Oral presentation (general)
Venue：onsite (Guangzhou, China / Kumamoto, Japan) and online
-
Cattle Face Identification with Ear Tags Using YOLOv5s Model International conference
Wai Hnin Eaindrar Mg and Thi Thi Zin
The Fifth International Symposium on Information and Knowledge Management (ISIKM2022) (onsite (Guangzhou, China / Kumamoto, Japan) and online) 2022.3.26 ICIC International
Event date： 2022.3.25 - 2022.3.26
Language：English Presentation type：Oral presentation (general)
Venue：onsite (Guangzhou, China / Kumamoto, Japan) and online
-
Black Cow Localization and Tracking with YOLOv5 and Deep SORT International conference
Cho Cho Aye, Thi Thi Zin, I. Kobayashi
The Fifth International Symposium on Information and Knowledge Management (ISIKM2022) (onsite (Guangzhou, China / Kumamoto, Japan) and online) 2022.3.26 ICIC International
Event date： 2022.3.25 - 2022.3.26
Language：English Presentation type：Oral presentation (general)
Venue：onsite (Guangzhou, China / Kumamoto, Japan) and online
Nowadays, precision livestock farming becomes popular fields in agriculture to get a high-quality product. Lacking monitoring livestock overtime can lead them to unexpectedly suffer from serious illness or even life-threatening. Therefore, individual livestock tracking systems can help to understand their actions. The proposed system is emphasized tracking individual black cows after performing cow detection. In this work, YOLOv5 (You Only Look Once) is used for detection and the Deep SORT (Simple Online Real-time Tracking) algorithm is utilized for tracking. In order to enhance the accuracy of the models, transfer learning is adopted. In the detection phase, the YOLOv5 model and appear feature model in Deep SORT were trained with a manually annotated dataset with help of transfer learning. The experimental results reveal that YOLOv5 got detection accuracy of 0.995 mAP@.5 and tracking with Deep SORT achieves the tracking accuracy of 99.4% in testing video-1 and 98.9% in testing video-2.
-
工場での作業の見える化 - 作業員のグループ識別及び追跡 - Invited
Thi Thi Zin
令和３年度先端技術研究開発促進・人材育成支援事業：IoT等先端技術利活用セミナー 2022.3.17
Event date： 2022.3.17
Language：Japanese Presentation type：Public lecture, seminar, tutorial, course, or other speech
-
Action Recognition System for Senior Citizens Using Depth Image Colorization International conference
Ye Htet, Thi Thi Zin, H. Tamura, K. Kondo, E. Chosa
The 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech 2022) (Osaka, Japan) 2022.3.8 IEEE Life Sciences Technical Community
Event date： 2022.3.7 - 2022.3.9
Language：English Presentation type：Oral presentation (general)
Venue：Osaka, Japan Country：Japan
This paper describes about the system which can be used at the care center for the purpose of elderly action recognition using depth camera. The depth image colorization is used for compression, visualization, and person detection process. The YOLOv5 (You Only Look Once) algorithm is used as object detector. The space-time features are extracted from depth sequences and they are recognized by linear SVM (Support Vector Machine) classifier. The random image sequences are generated for testing to recognize six actions. The results show that this system can detect the various actions with the average of 92% accuracy for different durations.
-
A Hybrid Approach: Image Processing Techniques and Deep Learning Method for Cow Detection and Tracking System International conference
Cho Cho Mar,Thi Thi Zin, I. Kobayashi, Y. Horii
The 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech 2022) (Osaka, Japan) 2022.3.8 IEEE Life Sciences Technical Community
Event date： 2022.3.7 - 2022.3.9
Language：English Presentation type：Oral presentation (general)
Venue：Osaka, Japan Country：Japan
Cow detection and tracking system plays an important role in cattle farming and diary community to reduce expenses and workload. This research presents how the conventional image processing techniques can be combined with deep learning concepts to establish cow detection and tracking system. Specifically, we first employ a Hybrid Task Cascade (HTC) instance segmentation network for cow detection. We then built the multiple objects tracking (MOT) algorithm utilizing location and appearance cues (color and CNN features) to carry out cow tracking process. To leverage the robustness of the system, we also considered the recent features from the previous tracked cow.
-
A Study on Worker Tracking Using Camera to Improve Work Efficiency in Factories International conference
I. Hidaka, S. Inoue, Thi Thi Zin
The 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech 2022) (Osaka, Japan) 2022.3.8 IEEE Life Sciences Technical Community
Event date： 2022.3.7 - 2022.3.9
Language：English Presentation type：Oral presentation (general)
Venue：Osaka, Japan Country：Japan
The population of Japan is declining every year. As the population declines, the number of employees in enterprises also decreases, and one of the issues for small and medium-sized enterprises is the decline in productivity due to a shortage of employees. It is necessary to improve work efficiency to compensate for the shortage of employees and the resulting decrease in productivity. If changes in the work process and unnecessary movements in the work process can be eliminated, it will lead to shortening of work time and improvement of work efficiency. In this paper, to improve work efficiency for factories, we aim to track the workers in the factory using a camera and show the trajectory of works.
-
A Study on Automatic Individual Identification of Wild Horses International conference
K. Shiiya, R. Yamada, Thi Thi Zin, I. Kobayashi
The 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech 2022) (Osaka, Japan) 2022.3.8 IEEE Life Sciences Technical Community
Event date： 2022.3.7 - 2022.3.9
Language：English Presentation type：Oral presentation (general)
Venue：Osaka, Japan Country：Japan
Wild horses called "Misaki-uma" are inhabited Cape Toi in the southern part of Miyazaki Prefecture in Japan. Although wild horse, it needs health care. It is a difficult task for aging management association members to monitor the vast habitat range of horses and to identify individual during routine management. In addition, the current methods of individual identification because of contact with horses or to requiring specialized knowledge, ordinary people cannot perform it. In this paper we propose a method for automatic individual identification of wild horses without contact using an RGB camera.
-
A Deep Learning-based solution to Cattle Region Extraction for Lameness Detection International conference
Su Myat Noe, Thi Thi Zin, Pyke Tin, I. Kobayashi
The 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech 2022) (Osaka, Japan) 2022.3.8 IEEE Life Sciences Technical Community
Event date： 2022.3.7 - 2022.3.9
Language：English Presentation type：Oral presentation (general)
Venue：Osaka, Japan Country：Japan
In precision livestock farming, lameness detection in cattle is particularly important for breeding management. The accurate detection of lameness is crucial for delivering effective and economical treatment and for preventing future diseases. The noticeable sign of lameness is that their speed of walking, arching their backs and drop their heads during walking. Here, we emphasis on lameness of dairy cattle by implementing the intelligent visual perception system on the laneways after milking process. Employing a deep learning technique of Mask-RCNN for cattle region detection and identification. The novelty of this work noticeably implies that deep learning instance segmentation could be effectively employed as a cattle region extraction from complex background prior to using identification and tracking.
-
An Intelligent Method for Detecting Lameness in Modern Dairy Industry International conference
Thi Thi Zin, Moe Zet Pwint, Su Myat Noe, I. Kobayashi
The 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech 2022) (Osaka, Japan) 2022.3.8 IEEE Life Sciences Technical Community
Event date： 2022.3.7 - 2022.3.9
Language：English Presentation type：Oral presentation (general)
Venue：Osaka, Japan Country：Japan
Lameness is one of the major welfare concerns in the modern dairy industry. In addition, lameness makes severe health and economic problems causing losses in milk production. Although there has a sizable amount of methods, it remains some worthwhile open problems. Therefore, in this paper, we propose an intelligent method for detecting the lameness of dairy cow by establishing a visual monitoring system on the laneways after milking process. We employ a technique of Mask-RCNN for cow region extraction and utilize features based on head bob patterns. Our real-life experimental results show that the proposed method has detection accuracy of 95.5% on cow’s region extraction and can classify 80% of the lameness levels correctly.
-
Individual Identification of Cow Using Image Processing Techniques International conference
Y. Kawagoe, Thi Thi Zin, I. Kobayashi
The 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech 2022) (Osaka, Japan) 2022.3.8 IEEE Life Sciences Technical Community
Event date： 2022.3.7 - 2022.3.9
Language：English Presentation type：Oral presentation (general)
Venue：Osaka, Japan Country：Japan
Cow identification has become important in recent years due to outbreaks of diseases such as bovine spongiform encephalopathy. Conventional identification methods available are not efficient, affordable, non-invasive, and cost- effective. Among them, some methods are based on biological markers, such as muzzle point matching and facial recognition. Facial images are the most common biometric characteristics used by humans to identify individuals, and they have received much attention. In this study, we used RGB camera to identify individual cow by their faces and confirmed the effectiveness of this method.
-
高度な画像処理技術やAI技術を活用した研究開発 Invited
Thi Thi Zin
ICT研究開発支援セミナーin九州 (Zoom及びYouTubeによるオンライン配信) 2022.2.4 九州総合通信局
Event date： 2022.2.4
Language：Japanese Presentation type：Public lecture, seminar, tutorial, course, or other speech
Venue：Zoom及びYouTubeによるオンライン配信 Country：Japan
総務省の戦略的情報通信研究開発推進事業(SCOPE)「地域ICT振興型研究開発」採択課題である「ICTを活用した牛のモニタリングシステムの開発」を中心に、「自立生活を支援するための高齢者24時間見守りシステムの開発」等の高度な画像処理技術やAI技術を活用した研究について紹介。
-
A Simple Random Walk Model for Dairy Cow Calving Time Prediction International conference
Thi Thi Zin, Pyke Tin, Pann Thinzar Seint, K. Sumi, I. Kobayashi, Y. Horii
2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) (Kyoto, Japan) 2021.10.14 IEEE Consumer Technology Society
Event date： 2021.10.12 - 2021.10.15
Language：English Presentation type：Oral presentation (general)
Venue：Kyoto, Japan Country：Japan
In this paper, we propose a simple random walk model to predict time of calving event occurs for a pregnant dairy cow. Dairy farmers and experts have well recognized that an accurate calving time prediction is quite important in modern smart dairy farming. To meet these demands, we consider the number of posture changes of a pregnant cow during a few days before the expected dates as a random walk to predict the time at which the calving event occurs. For validation, we show some experimental results by using real life data collected from a large dairy farm in Japan.
-
Evaluation of the severity of tremor based on each signal acquired from the displacement of the hand movements International conference
T. Hayashida, Thi Thi Zin, K. Sakai, H. Mochizuki
2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) (Kyoto, Japan) 2021.10.14 IEEE Consumer Technology Society
Event date： 2021.10.12 - 2021.10.15
Language：English Presentation type：Oral presentation (general)
Venue：Kyoto, Japan Country：Japan
Discrepancy of exam findings at the same patient make it difficult to ascertain chronological change in the disease and the efficacy of the medicine. Quantitative evaluation of severity is important for improving the discrepancy. In this study, we examine the efficacy of quantitative evaluation of tremor when using single camera. Recording the hand movements of tremor with single camera, and the displacement, velocity, and acceleration signals are acquired using the hand shift between two adjacent video frames. Quantitative evaluation of tremor is performed based on features obtained from each signal. According to the validation results, our method using single camera is possible to classify with an accuracy of up to 82.6%.
-
Cattle Region Extraction using Image Processing Technology International conference
Y. Motomura, Thi Thi Zin, Y. Horii
2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) (Kyoto, Japan) 2021.10.14 IEEE Consumer Technology Society
Event date： 2021.10.12 - 2021.10.15
Language：English Presentation type：Oral presentation (general)
Venue：Kyoto, Japan Country：Japan
In recent years, the number of dairy and beef cattle farms has been decreasing, while the number of cattle and the number of cattle per farm have been increasing, so systems for automatically monitoring cattle have been actively introduced. However, most of them are contact type, which causes physical or mental stress to the cows and is costly when the equipment is damaged. Therefore, in this research, we proposed a method for extracting the approximate shape of cattle using a non-contact 360-degree camera to reduce the burden on livestock farmers and cattle, and confirmed its effectiveness through experiments.