111 research outputs found
Hypergraph based visual categorization and segmentation
No full textThis dissertation explores original techniques for the construction of hypergraph models for computer vision applications. A hypergraph is a generalization of a pairwise simple graph, where an edge can connect any number of vertices. The expressive power of the hypergraph models places a special emphasis on the relationship among three or more objects, which has
made hypergraphs better models of choice in a lot of problems. This is in sharp contrast with the more conventional graph representation of visual patterns where only pairwise connectivity between objects is described. The contribution of this thesis is fourfold: (i) For the first time the advantage of the hypergraph neighborhood structure is analyzed. We argue that the summarized local grouping information contained in hypergraphs causes an ‘averaging’ effect which is beneficial to the clustering problems, just as local image smoothing may be beneficial to the image segmentation task. (ii) We discuss how to build hypergraph incidence structures and how to solve the related unsupervised and semi-supervised problems for three different computer vision scenarios:
video object segmentation, unsupervised image categorization and image retrieval. We compare
our algorithms with state-of-the-art methods and the effectiveness of the proposed methods is demonstrated by extensive experimentation on various datasets. (iii) For the application of image retrieval, we propose a novel hypergraph model — probabilistic
hypergraph to exploit the structure of the data manifold by considering not only the local grouping information, but also the similarities between vertices in hyperedges. (iv) In all three applications mentioned above, we conduct an in depth comparison between
simple graph and hypergraph based algorithms, which is also beneficial to other computer vision applications.Ph.D.Includes bibliographical referencesIncludes vitaby Yuchi Huan
A multi-step heart rate prediction method based on physical activity using Adams-Bashforth technique
No full text[[alternative]]Parents' Choice of Transregional Public Kindergarten: A Case Study of Yuchi Township
No full text[[abstract]] 早期幼教品質參差不齊，不管在教學、環境、師資或行政體制…等等方面，充斥著各自發展的亂象，政府為因應幼托整合方案，從2005年起開始試辦國幼班，屬實驗性質，至2011年六月，立法院三讀通過《幼兒教育及照顧法》法案，宣告2012年元旦開始實施，將幼稚園及托兒所整合為幼兒園，招收兩歲到六歲的學齡前幼兒；在課程方面擬定新課程綱要等配套措施，由教育部監督管理。 縱觀幼教改革從2005年至今2019年，歷經十五年的時間，其初為改善部分因社會不利條件或限制（偏遠地區交通不利）無法順利就學之幼童的學習機會，在離島三縣三鄉及原住民鄉鎮市，國小部內增設國幼班（招收4-6）歲學齡前幼兒，並招考代理幼教師，進駐學校，興建幼兒園教室、充實及改善教學環境，當時魚池鄉（南投縣原住民三鄉-魚池鄉、仁愛鄉、信義鄉）也在國幼班的輔導之下逐漸成長。 其後在2012年開始實施《幼兒教育及照顧法》法案，改制為幼兒園，魚池鄉共有八所國小附設幼兒園，近年來由於少子化的衝擊，加上各公私立幼兒園林立，招生變得越來越困難。 統計調查魚池鄉八所公立附設幼兒園，從2015年至2019年幼生人數的變動，可以發現各所公立附設幼兒園幼生的就學率幾乎都是下降的情況，但唯獨東光國小附設幼兒園的幼生就學率有不降反增的狀況，到底是甚麼原因？ 作者所服務的學校，為魚池鄉東光國小附設幼兒園，其中有三分之二是越區就讀的孩子，本研究採用問卷調查法及質性研究法，從文獻探討、調查表及訪談的過程中，了解影響幼兒園家長選擇學校時的資訊來源及考量因素，希望透過家長的角度，了解家長在為幼兒選擇學校時，真正在意的需求，根據訪談結果作綜合歸納統整,並分析與討論，最後提出具體的結論與建議，提供給予幼兒園經營者及教師作為改善之參考。[[abstract]] The quality of early childhood education varies, regardless of the teaching, environment, teachers or administrative system, etc., which are full of chaos in their respective developments. In response to the integration of child care, the government has started piloting national child care classes since 2005. The nature of the experiment. By the month of 2011, the three courts passed the Early Childhood Education and Care Act, announcing that it will be implemented on New Year 's Day 2012, integrating kindergartens and nurseries into kindergartens, and enrolling pre-school children from 2 to 6 years old. The new curriculum outline and other supporting measures are drawn up and supervised by the Ministry of Education. Throughout the fifteen years since the reform of preschool education from 2005 to 2019, at the beginning, in order to improve the learning opportunities of some children who could not go to school due to social disadvantages or restrictions (unfavorable transportation in remote areas), they were In townships and aboriginal towns and cities, additional elementary school children (4-6) years old preschool children were recruited in the elementary school, and enrolled acting young teachers, settled in schools, built kindergarten classrooms, enriched and improved the teaching environment. At that time, Yuchi Township (Nantou County) Indigenous three townships-Yuchi Township, Renai Township, and Xinyi Township) have also gradually grown up under the guidance of the National Youth Class. Since 2012, the ＂Early Childhood Education and Care Act＂ has been implemented into a kindergarten. There are eight kindergartens attached to Yuchi Township. In recent years, due to the impact of declining birthrates, and the addition of public and private kindergartens, the enrollment has become more and more It's getting harder. According to the statistical survey of the eight public kindergartens in Yuchi Township, the change in the number of infants from 2015 to 2019 can be found that the enrollment rate of young public kindergartens in all public kindergartens has almost declined, but only the pupil enrollment rate of the kindergarten attached to Dongguang Elementary School What is the reason why there is no decline but increase? The school served by the author is a kindergarten attached to Dongguang Elementary School in Yuchi Township, two-thirds of which are children from Yue District. This research uses questionnaire survey and qualitative research methods, from literature discussion, questionnaire and interview process In order to understand the sources of information and considerations that affect kindergarten parents when choosing a school. From the perspective of understanding the needs that parents really care about when choosing a school for young children, make a comprehensive summary based on the interview results, analyze and discuss, and finally put forward specific conclusions and recommendations to provide kindergarten operators and teachers as a reference for improvement
Chemistry Lab Automation via Constrained Task and Motion Planning
Full text linkChemists need to perform many laborious and time-consuming experiments in the
lab to discover and understand the properties of new materials. To support and
accelerate this process, we propose a robot framework for manipulation that
autonomously performs chemistry experiments. Our framework receives high-level
abstract descriptions of chemistry experiments, perceives the lab workspace,
and autonomously plans multi-step actions and motions. The robot interacts with
a wide range of lab equipment and executes the generated plans. A key component
of our method is constrained task and motion planning using PDDLStream solvers.
Preventing collisions and spillage is done by introducing a constrained motion
planner. Our planning framework can conduct different experiments employing
implemented actions and lab tools. We demonstrate the utility of our framework
on pouring skills for various materials and two fundamental chemical
experiments for materials synthesis: solubility and recrystallization.Comment: Equal author contribution from Naruki Yoshikawa, Andrew Zou Li,
Kourosh Darvish, Yuchi Zhao and Haoping X
Selected Tales from Georgia Indian Tribes with Suggested Application in Junior High School English and Social Studies Curricula
Full text linkAfter finding several references which suggest that there was a need to include Native American Studies in schools, the author of this report decided to create, Cherokee, Creek, and Yuchi tales with Suggested Activities ; a resource supplement that would be of help to teachers of junior high school English and/or social studies in Columbus.
The completed resource supplement on Georgia Indian folk-tales include tales from three tribes, Cherokee, Creek, and Yuchi, selected because of their importance and proximity to Columbus. Twenty-three tales were chosen based on criteria established by the author.
The resulting collection of stories and instructional suggestions includes tales, a glossary, goals, objectives, and suggested activities for use with the book. The readability level of all tales was checked using the Fry formula. Afterwards, the included tales were rewritten on a fourth/fifth grade level and reevaluated by the Fry formula. The objectives conform to the cognitive, affective, and psychomotor domains. These activities are not considered to be all-encompassing--they offer a point from which a teacher may begin.
Appendices following the supplement consist of the classification of tales as to tribe, type of tale, and original readability level of each. A list of the sources of the tales follows. Also included in the appendices are maps which show the general area in which the Cherokee and Creek tribes resided and pinpoint specific locations referred to in some of the tales
[[alternative]]The Success Rate of Young Elite Badminton Players of Anticipating Overhead Backhand Shots of Different Skill Leveled Opponents
No full text[[abstract]]This research is to examine the accuracy of anticipation of points of landing of an overhead backhand shot based on visual pre-cues of different stages of the shot motion given to the player, and furthermore discover the crucial cues of different overhead backhand shots. The subjects of this research(N=52) are young elite athletes of the age of 16.5±1.20. With three female athletes with different skill levels of professional training to provide the footage of a series of overhead backhand shots. With the use of a digital video camera and post production to produce three stages of visual pre-cueing and a total of 180 shots to use in a temporal occlusion paradigm. The subjects are to decide on the landing point within three seconds of viewing the images. The rate of accurate identification was calculated using paired T-test as the method of statistical analysis, and showed a clear level of 0.5.
The results show:
1.Through the different stages and increase of visual pre-cues, the success rate of identifying landing points of the high level and low level athlete greatly increased. The difference in successful identification rates of the 3 stages of motion all achieve clear level.
2.In stage one of visual pre-cueing, the difference in success rate of identifying landing points did not achieve a clear level and were close to identical among the three skill levels. In stage 2, the success rate of identifying landing points between high skilled athlete and middle skilled athlete were clearly different. In stage 3 of visual pre-cues, all three skill levels had different success rates of identifying landing points.
3.The difference in success rates of identifying landing points between high skilled and middle skilled athletes were only similar for straight drop shot and cross-court drop shot, all other shots achieved a clear level of difference. The prediction of the four shots of the low skilled level athlete all achieved a clear level of difference.
Energy harvesting for wearable and portable devices
No full textThe aim of this chapter to provide an overview of the different techniques that can be used to harvest energy for portable and wearable electronic devices. These technologies rely on converting ambient energy, or energy from the body into useful electricity. We discuss each technology‘s working principles, as well as its merits for satisfying the power demands of wearable electronic devices
Renewing Steiner Teacher Education: A Conversation With Florian Osswald
No full textThis article contains an interview between the author and Florian Osswald, one of the co-leaders of the Steiner education movement worldwide, held at the end of the four-day Asian Steiner Teacher Educators’ Conference in Yuchi, Taiwan in February, 2020. It focuses on the work of the International Teacher Education Project (ITEP), an initiative set up by Osswald in 2017; the aim of ITEP is the renewal of Steiner teacher education worldwide, through a process of long-term engagement with teacher educators around the world. The interview covers the consultative process Osswald has chosen for ITEP and indicates how a century-old, globally distributed, spiritually based educational movement is looking to maintain that relevance. It gives insight into the current status of Steiner teacher education and some indications of directions in which it might develop further
Thickness of the Seasonal Deposits at the Martian North Polar Region From Shadow Variations of Fallen Ice Blocks
No full textAbstractThe seasonal deposition and sublimation of CO2 constitute a major element in the Martian volatile cycle. Here, we propose to use the shadow variations of the ice blocks at the foot of the steep scarps of the North Polar Layered Deposits (NPLD) to infer the vertical evolution of the seasonal deposits. We conduct an experiment at a steep scarp centered at (85.0°N, 151.5°E). We assume that no snowfall remains on top of the selected ice blocks, the frost ice layer is homogeneous around the ice blocks and their surroundings, and no significant moating is present. We show that the average thickness of the seasonal deposits due to snowfalls in Mars Year 31 is 0.97 ± 0.13 m at Ls = 350.7° in late winter. The large depth measured makes us wonder if snowfalls are more frequent and violent than previously thought. Meanwhile, we show that the average frost thickness in Mars Year 31 reaches 0.64 ± 0.18 m at Ls = 350.7° in late winter. Combined, the total thickness of the seasonal cover in Mars Year 31 reaches 1.63 ± 0.22 m at Ls = 350.7° in late winter, continuously decreases to 0.45 ± 0.06 m at Ls = 42.8° in middle spring and 0.06 ± 0.05 m at Ls = 69.6° in late spring. These estimates are up to 0.8 m lower than the existing Mars Orbiter Laser Altimeter results during the spring. Meanwhile, we observe that snow in the very early spring of Mars Year 36 can be 0.36 ± 0.13 m thicker than that in Mars Year 31. This study demonstrates the dynamics of the Martian climate and emphasizes the importance of its long‐term monitoring.Plain Language Summary: Like Earth, Mars also has seasons. Up to one third of the atmospheric CO2 annually exchanges with the polar surface through seasonal deposition/sublimation processes. Deposition can be either atmospheric precipitation as snowfall or direct surface condensation as frost. At the steep scarps of the North Polar Layered Deposits (NPLD), fractured ice fragments can detach and fall to form ice blocks. We propose to use variations in the shadows of these ice blocks, observed in the High Resolution Imaging Science Experiment images, to infer the thickness evolution of the seasonal deposits. We make reasonable assumptions about the distribution of snowfall and frost around the ice blocks and their surroundings, which allow us to separately measure the thickness of snowfall and frost. Meanwhile, we introduce a novel approach that allows us to estimate the thickness of the seasonal deposits during late winter and early spring when image quality is insufficient. This approach also enables us to peer into the interannual thickness variations of snowfall. We carry out a successful experiment at a scarp centered at (85.0°N, 151.5°E). The obtained thickness measurements demonstrate the dynamics of the Martian volatile cycling and can be used to constrain the Martian climate models.Key Points:
We propose to examine the shadow variations of the ice blocks at the Martian polar region to infer the thickness of the seasonal deposits
Maximum thickness of the seasonal deposits at the study scarp in MY31 is 1.63 ± 0.22 m to which snowfalls contribute 0.97 ± 0.13 m
Seasonal deposits at the study scarp are up to 0.8 m shallower than previous measurements during spring
HX, LML, and PJGhttps://doi.org/10.17189/1520303https://doi.org/10.17632/5yy475dbry.1https://doi.org/10.17632/x953mzxxvv.1https://doi.org/10.17189/1520101http://www.msss.com/moc_gallery/200
A Research of Heart Rate Prediction Model Based on Evolutionary Neural Network
No full textHeart rate (HR) signal analysis is widely used in the medicine and medical research area. Physical activities (PA) are commonly recognized to greatly affect the changes of heart rate. A method of Evolutionary Neural Network -- Neuroevolution of Augmenting Topologies (NEAT) is used to build a PA-based HR predictor model. Through special coding, crossover and mutation operator, NEAT can implement network topology and connectivity weights evolution simultaneously. The common problem in evolutionary neural network, like competing conventions, how to protect the new innovation are effectively solved. The experimental results demonstrated the application potential of the approach.No Full Tex
- …
