Chung Cheng University Institutional Repository
Not a member yet
889 research outputs found
Sort by
基於相位移結構光掃描技術應用於人體表面快速重建;A Phase-Shifting Structure Light Based Scanner Applied to Fast Reconstruct Body Surface
No full text[[abstract]]本研究目的為開發一套非侵入式的呼吸追蹤系統，追蹤患者在進行放射治療時的腫瘤位置，由於腫瘤會隨著呼吸運動在體腔內位移，因此患者需要在正確的時間接受放射線，避免放射治療時攻擊到正常的細胞組織。本研究利用相位移結構光掃描技術重建人體表面並使用貼附於人體皮膚的標記點(Marker)表示呼吸狀態，Marker會在手術前投影到4D CT的模型上產生虛擬Marker。在呼吸追蹤時，利用人體皮膚上的真實Marker與虛擬Marker進行比對找到最接近的CT模型，藉此得知患者的呼吸狀態以及腫瘤位置是否在預定照射放射線的位置。本研究所開發的系統在讀取手術規劃的放射角度後可以運用於手術模擬以及教育訓練，此外本研究利用馬達機構實驗動態追蹤Marker的準確性，追蹤結果與理想波形有0.9994的正相關性。
The purpose of this research is to develop a non-invasive respiratory tracking system to track the position of the tumor during radiation therapy. Because the tumor would move with respiratory movement in the body, the patient needs to receive the radiation at the correct time window to prevent the normal cell tissue from getting hurt.We use the phase-shift technology to reconstruct the skin of the human body with markers attached to the human body. The markers will be projected to the STL skin model from 4D CT to generate virtual markers. While doing respiratory tracking, the system will use the markers on the body to match the virtual markers and find the closest STL model. Such corresponding STL model can be used to find the tumor at the right position for radiation projection. Our system can read the angles of radiation of surgical planning and use that information in surgical simulation and educational training. Furthermore, we use a motor mechanism to assess the accuracy of the marker tracking. The result shows that the PCC correlation of the tracking was found to be 0.9994 correlation with the ideal waveform
影像處理與機器學習於鏟花加工面之研究;A Research of Scraping Surface Inspection with Image Processing and Machine Learning
No full text[[abstract]] 台灣工具機產業出口面臨複雜的國際競爭，唯有提升產業技能才能確保台灣的競爭力。鏟花針對工具機而言非常重要，目前文獻都針對鏟花做影像處理，目標都是評定鏟花的承斑點數及面積比，或研發自動鏟花機構來取代人工鏟花，達到自動化工廠之目的。本研究利用影像處理與機器學習之方法來研究鏟花加工面，經由影像處理可以去除不必要的雜訊並擷取所需要的特徵，再利用機器學習方式運算擷取後之特徵，而機器學習的部分以「倒傳遞類神經網路」的方法進行研究，以大數據分析其工件樣本再加以訓練，經由良好網路輸出可信賴的鏟花數據，本研究影像處理經由機器學習後的正確率無論是分類或PPI的預測皆可達90%。
Scraping technique is one of the key point techniques in machine tool industry. The quality of scraping are mostly inspected manually or judged automatically using image processing methods. Since the scraping sizes, shapes and patterns of work pieces are different from another, manual judgement can be subjective. In addition, different image processing methods can lead to different inspect results. Thus, in addition to image processing methods, machine learning is applied in this study to make the inspection results more reliable. The experimental results show that with proposed method, both of the correct scraping classification rate and correct PPI prediction rate by image processing followed by the machine learning method can reach up to 90 %
新興及再浮現傳染病對人類安全之衝擊：以2015年我國南台灣登革熱為例;The Impact of the Emerging and Re-emerging Infectious Diseases on Human Security：A Case Study of Southern Taiwan Dengue Virus in 2015
No full text[[abstract]]新興及再浮現傳染病問題隨著國際觀光貿易旅遊的往來頻繁及全球氣候變遷等因素，使得疾病之擴散更加迅速，不僅成為我國社會所需面臨的課題，也是全球性的安全危機。因此，新興及再浮現傳染病之防治，近年來已經成為公共衛生的重要目標。 隨著國際情勢的轉變以及經貿全球化導致各國依存度增加，國際上對於安全觀的思考不斷演進，也隨之發展出許多新的安全概念；健康安全就是人類安全的七大安全類型之一，人類安全議題愈來愈不容忽視，舉凡日益頻繁的非法入境、跨國犯罪、環境汙染、種族歧視與新興傳染病爆發事件等等，都處處影響著人類與國家的發展。因此，新興及再浮現傳染病對於全球健康安全的重要性是名副其實的。 本研究除了探討新興及再浮現傳染病之外，並針對全球性國際組織對於此類型傳染病的因應之道詳細分析解說，並提出我國南台灣於2015年爆發登革熱這個案例加以說明，就其對我國社會及醫療體系的所造成衝擊分別闡述，並期望透過國際社會的機制與應對措施，將有效且完善地因應新興及再浮現傳染病的全球威脅與挑戰。
With the frequent travel of international tourism and tourism and global climate change and other factors, making the spread of the Emerging and Re-emerging infectious diseases more rapid. Not only become our society needs to face the subject, but also a global security crisis. Therefore, the Emerging and Re-emerging infectious diseases prevention and control, has become an important goal of public health in recent years. With the international situation changes and economic globalization led to increased dependence on countries, international thinking on the concept of security evolved, also developed a number of new security concepts. Health Security is one of the seven safety types of Human Security, Human Security issues can't be ignored. As a result, in addition to exploring Emerging and Re-emerging infectious diseases and in the context of detailed analysis and analysis of global international organizations' response to this type of infectious disease and the case study of southern Taiwan dengue virus in 2015. Description caused by the impact of respectively on our health system and southern Taiwan. And I hope that through the mechanisms and responses of the international community and will effectively and perfectly address the global threats and challenges of Emerging and Re-emerging infectious diseases
李鴻章在自強運動中角色之研究;A Study of the Role of Li Hongzhang in the Self-Strengthening Movement.
No full text[[abstract]]1840年6月(道光二十年五月)，英國對清廷發動第一次鴉片戰爭，清廷不敵英軍的戰艦大炮宣告戰敗，兩國於1842年8月29日(道光二十二年七月二十四日)簽訂了中國第一個不平等條約－《南京條約》。就此開啟了清朝被西方列強蠶食鯨吞的命運。1856年(咸豐六年)，英法聯軍再次聯合攻打中國，中國戰敗，此時的清廷開始出現了「師夷之長技以制夷」的一連串軍事、商務、教育、民生工業等自強改革運動的思想。 其中，李鴻章自1862年(同治元年)抵達上海後，便開始籌畫並著手洋務運動的事業，且繼曾國藩接任直隸總督之後，直到甲午戰爭期間，一直始終其事，從事於全國性自強建設的事業。從最初的軍事工業建設及新式教育建設，到後來的商務民生工業建設，幾乎都由李鴻章所創辦或策畫。雖然最後自強運動以失敗告終，但其歷史意義和重要性仍是不可抹滅的。
In June 1840, the United Kingdom launched the first opium war against the Qing China, who lost to British warships and artillery, and declared defeat. This began the erosion of the Qing China by Western powers. In 1856, a British and French military alliance once again carried out joint attacks on China; with China’s defeat, the Qing Court began to exhibit new ways of thinking on a series of military, commercial, education, public welfare, and industrial matters, as a part of the “Self-Strengthening Movement.” After arriving in Shanghai in 1862, Li Hongzhang began to plan and execute the political causes of the Self-Strengthening Movement. From the time that Li succeeded Zeng Guofan as the Viceroy of Zhili, to the period of the First Sino-Japanese War, Li was occupied with establishing the Movement on a national scale. Nearly all of the reforms from military, industrial, and educational modernization to subsequent commercial and public welfare reforms were planned or implemented by Li Hongzhang. Although the Self-Strengthening Movement ended in failure, its historical significance and importance remain indelible
美日在兩岸關係之角色分析:美中競合的研究觀點;Analyses of the relationship of American and Japan under the cross-strait relations:the viewpoint of China and American competing
No full text[[abstract]]近年來中國崛起的速度愈來愈快，尤其經濟上的成長更為明顯。美國意識到中國崛起後，害怕在東亞的利益會遭受中國的威脅，因此美國國務卿柯林頓在2010年7月表示美國將重返亞洲，而未來十年美國會把經濟、外交的重點放在亞洲地區。面臨全球化的時代，美國和中國處於既競爭亦合作的關係，美國和中國的發展互動也將影響到兩岸關係的發展。兩岸關係的發展從軍事衝突階段，和平對峙階段，民間交流階段一路演變下來，雖然說兩岸關係的發展愈來愈和平，但中國大陸在國際上時常以打壓的方式，而使台灣備感壓力。本研究試從美中競合來分析。美國重返亞太後，與中國的競爭及合作的關係下，對兩岸關係發展會有怎樣的變化，兩岸關係會掉入困境或有新的契機出現。本研究將採用文獻分析研究法，結合?岸與國際現勢，透過新現實主義理論來探討美國在亞太戰略布局，及中國如何回應，並從中探究對台海問題有甚麼樣的影響，並對台灣未來在兩岸關係發展做出建議。
The speed of China's rise has become faster and faster in recent years, especially economic growth. United Stated realizes that China's rise, and they afraid the benefit of East Asia threaten by China. Therefore, in the July, 2010, United States Secretary of State, Clinton’s mentioned that United Stated would pivot to Asia. In addition, they will focus on economic and diplomatic interests in the Asian region. In the generation of globalization, United Stated and China would stay not only competitive but also cooperative relationship. The interaction of United States and China will influence development of cross-strait relationship.The development of cross-Strait relationship begin from military affairs to peace confrontation, and people-to-people contact. Although the developing cross-strait relationship getting more and more peaceful, Taiwan still feel stressful because China keep bashing on Taiwan in the International Communities. The research is used to analyze US-China Relations. What will happen to the development of cross-strait relations in the US-China Relations after United States pivot to Asia. Cross-Strait relationship will get into trouble or have new opportunity. The research will adopt documentary analysis, to combine Cross-Strait and international relations to discuss the strategy in Asia-Pacific of United States through Neo-Realism. According to the response of China to discuss Taiwan Strait problem, and to make recommendations to Taiwan the development of Cross-Strait relationship in the future
過量零值資料上的迴歸樹方法;
No full text[[abstract]]在分析計數型資料時一般會利用普瓦松迴歸模型，但實際搜集到的資料若有著過度散佈或是零值過多的情形，推論的結果容易因違背模型假設而產生誤導，此時若使用普瓦松迴歸樹分析資料，則可能會導致方法本身在選取切割變數時失去挑選正確切割變數的能力或是存在選取偏差，針對這些問題，本篇文章提出 tezc 迴歸樹方法，並探討此方法選取正確切割變數的能力與穩定性，同時與 GUIDE 和 MOB 迴歸樹做比較，我們發現在大部份情況下 tezc 皆優於其他兩種方法
具電源網路壓降感知之高效率直流對直流降壓型電源轉換器設計;Design of Power Network IR-Drop-Aware High Efficiency DC-DC Buck Converter
No full text[[abstract]]隨著先進製程的蓬勃發展，電壓壓降損耗在現代晶片製作中逐漸變成不可忽視的議題存在，隨著電晶體的尺寸及導線寬度變窄，電壓壓降損耗變得越來越嚴重。為了找出在電源分配網路中電壓壓降損耗最嚴重的路徑，設計者通常要從模擬結果中檢查所有可能會發生電壓壓降損耗的導通路徑，而現代多功能系統晶片的輸入端數量皆是百根腳位以上，這對設計者而言將是耗費數日才能完成的初步驗證，如果考量到多功能系統晶片中的電路區塊之動態負載變化及判定最嚴重之路徑準確度，那將會耗費許多人力與成本。因此在本論文中提出一個可應用在多功能晶片系統的具電源網路壓降感知之高效率直流對直流降壓型電源轉換器，利用混合訊號電路來改善因供應電壓在傳遞過程因傳遞路徑之等效電阻而導致電壓壓降損耗，進而讓後端電路效能穩定。此外，本架構提供額外技術，可根據多功能晶片系統的電流需求狀況進行供應排序智慧化，進而增加整體電路效能。 本論文使用台積電TSMC 0.18μm 1P6M 3.3V Mixed Signal製程製作，實驗結果顯示本論文提出的電源網路壓降感知機制能提升(所有電路區塊電壓壓降損耗總和)/N之電壓值(N=電路個數)，並有效的減輕電源分配網路成本。關鍵字：電源管理晶片、電源轉換器、直流轉直流電源轉換器、電源網路壓降感知、高轉換效率
Voltage drops are one of the most stringent problems in modern integrated circuit implementation, which is exacerbated by the decreasing transistor sizes and interconnect line widths. In order to find the worst case voltage drop that a power net of a design might suffer, the designer would have to check the voltage drops that occur from the simulation of all possible input vector pairs of a design. However, modern IC that have hundreds of inputs is difficult to simulate rapidly. It’s will waste much time and resource to finish it. Moreover, if we need to ensure accurate and consider all dynamic loading transient in blocks, it’s become more difficult. Therefore, we present a switching buck converter with technique of reducing IR-drop, which consist of mixed-signal circuit and traditional converter. We can improve performance of blocks by reducing IR-drop and make blocks stable. In this paper, the test chip is designed and fabricated with TSMC 0.18μm 1P6M 3.3V Mixed Signal process technology. Experiential results shows that this test chip can reducing (sum of all blocks IR-drop)/N (N = number of blocks) and reduce the cost of power distribution network.Key word: Power Management IC, Power Converter, DC-DC Converter, Power Network IR-Drop-Aware, High Power Efficienc
全積體互補式金氧半導體功率放大器;Fully Integrated Complementary Metal-Oxide-Semiconductor Power Amplifier
No full text[[abstract]]本篇論文目的在設計一符合5G系統規格之射頻功率放大器，操作頻帶為Ku頻帶。以TSMC 0.18 μm CMOS製程進行製作。電路輸出端採用架構為Aoki提出之主動分佈式變壓器達到功率結合與輸出匹配之目的。為符合系統需求，配合混波器之輸出，設計放大器輸入為差動輸入。放大器輸出以混波器輸出 -7 dBm能量作為設計參考。電路結構為三級放大器串接，預期達到增益20 dB、飽和輸出功率23 dBm之電路規格。電路輸出級與驅動級電晶體皆採用並聯疊接架構，整體電路共使用40顆電晶體。共閘極組態電晶體汲極端與閘極端偏壓皆為3.6 V、共源極組態電晶體閘極端偏壓為1.6 V。由於設計考量不周，造成模擬與量測差異。本篇論文共製作三顆功率放大器。透過模擬方式，找出電路問題發生原因。原因分別為線段佈局造成阻抗偏移、元件操作超越製程規範之功率以及接地路徑佈局不當造成電路振盪。考量電路熱效應、製程變異與量測環境等因素，重新模擬，達到模擬與量測結果吻合。
This study is devoted to the design of CMOS power amplifier used for the fifth communication system. A power amplifier was implemented in TSMC 0.18 μm CMOS technology. Amplifier uses transformer for output matching. Transformer combines four push-pull cells for high power exporting. The principle of distributed active transformer is power transfer from primary inductor to secondary inductor. Power transformation capability is determined by inductor’s quality factor and coupling coefficient. In order to meet the requirement of 23 dBm saturation output and 20 dB gain at least, two driver stages and one power stage are used in this study. The measured results are different from those by post-layout simulations. The possible root causes, such as imperfect RF choke and thermal effect were discussed in this thesis
利用模糊語意影像特徵表示法及FPGA平台之車內人臉偵測與定位研究;A Face Detection and Localization System for Driver Using Semantic-based Vague Image Representation and a FPGA Platform
No full text[[abstract]]近幾年來，台灣車輛肇事率逐年攀升，其事故發生原因多為駕駛未注意前方車輛行駛狀況所致。因此，近年來各大車廠陸續開發出可事先偵測並提醒駕駛者是否專心開車的預警系統，以期有效降低此類交通事故的發生。其中，以利用人臉來判斷駕駛之專注情況最值得研究。然而，由於車內空間有限，導致人臉偵測之運算能力也隨之受限。為此，本研究專注於開發一套低體積、低功耗之人臉追蹤系統，以便未來對人臉姿態及表情做進一步之分析。在本研究計畫中，系統首先利用數位相機取得人臉影像，接著採用模糊語意影像特徵表達法(Semantic-based Vague Image Representation, SVIR) 來完成人臉特徵辨識及追蹤。此演算法之特色在於以更少的運算資源，就能夠有效地達到人臉追蹤之目的。而精簡之演算過程，更能幫助本研究實現在系統單晶片之上，進而達到體積小、低功耗以及即時處理之效能。在晶片設計方面，則是採用現場可程式邏輯閘陣列(Field Programmable Gate Array, FPGA)晶片做為測試開發平台，以符合體積小與低開發成本之目的，以期能讓此系統未來能夠更加普及化。最後，在本論文的實驗結果中，本系統設計除了達成人臉偵測與定位之目的，系統資源使用量也僅占晶片運算資源的6.24%，消耗286.83 mW的功率。其功耗少於傳統偵測法使用的高耗能處理器，不但符合先前的研究目標，精簡化的運算過程也使得本系統能在低體積且低性能的晶片上達成高解析度和即時處理。因此，本研究的新式設計有助於未來基於嵌入式裝置的人臉偵測系統之發展。
The traffic accident in Taiwan has been gradually increased in recent years by distraction to vehicle in front. To improve this issue, many car manufacturers have developed various assistant systems to warn driver before accident. However, the lower computing power constrained by narrow space in car is a critical problem for developing an intelligent driving assistant system. For this, this research focuses on the feasibility for designing a compact, low-power, and low-cost face tracking for driver, which is helpful to monitor driver’s condition in real-time.In this research, a digital camera is employed to collect driver’s face. Meanwhile, a novel and concise algorithm of semantics-based vague image representation (SVIR) is also adopted to abstract driver’s face features in order to implement entire system on a single chip. Here a field-programmable gate array (FPFA) developing platform is working as a test bed in which it demonstrates the feasibility for a low-power and low-cost face tracking system.According to the experimental results, the proposed system has a lower power consumption at 286.83mW with only 6.24% hardware resource usage of chip. With this design, driver’s face can be effectively tracked during the scanning of image pixels. Such real-time performance is workable and promising for drive’s assistant system in the future
應用於卷積神經網路文字辨識之視覺重心前處理與雲端字典更正技術;Visual Center Preprocessing and Cloud Dictionary Correction Technique Applied to Text Recognition with Convolutional Neural Networks
No full text[[abstract]]「文字」是人類歷經好幾個世紀逐漸演化出來，是人與人之間溝通的符號，隨著影像辨識技術逐漸成熟，在自然場景的影像中偵測、辨識文字的辨識率已相當高，其中又以深度學習最為卓越，在深度學習中的卷積神經網路(Convolutional Neural Networks, CNN)近幾年來普遍被應用在文字偵測、辨識，但辨識率高的代價則是其運算複雜造成耗時過久，故此研究目標為利用影像前處理減少耗時以及雲端字典對詞彙校正。 首先，尋找文字出沒區域及顯著文字優先來得到視覺重心，藉此縮小辨識範圍來降低使用卷積神經網路辨識出字元之耗時，並在文字偵測部分由最大穩定極值區域 (Maximally Stable Extremal Regions, MSER)取代；再者，以往辨識出來的文字串通常是與自己建的詞彙庫比對得出最相近之單字，但此舉受限於詞彙庫，故以雲端字典取代得到較廣辨識能力以及使辨識容錯上升。
Text is the evolution of humanity after many centuries and the symbol of communication between people. As the image recognition technology matures, the recognition rate of the text in the natural scene is accurate. In many technologies, the deep learning is the best. Convolutional neural network in deep learning has been widely used in text detection and recognition in recent years. However, convolutional neural networks are computationally complex and time-consuming. Our research goals are reduce time-consuming by image preprocessing and word correction by using cloud dictionary. First, searching text areas and significant text priority are used to search visual centers. Visual centers are used to reduce the time-consuming of character recognition with convolution neural networks. The text detection uses MSERs to avoid time-consuming of the sliding window method. Second, the previous people are usually to obtain correct word by searching their own dictionaries. We use cloud dictionary for more efficient recognition and greater fault tolerance