Rendering technology has many applications in diverse fields, for example, digital archives, multimedia entertainment, inverse engineering, and even medical science. Beside the acquirement of accurate 3D geometry, how to get the best fitting feature to reconstruct the lighting reflectance of the object from different viewing direction is a critical problem. Read more


Digital matting technique are extensively applies to the compositing of the movies and animations, but it needs intensive artificial operations and hi-priced cost. Besides, the extraction of the foreground objects with fuzzy boundary, such as the hair is not an easy work. Besides, the unity of light state of foreground objects and the new background, also an important issue. Read more


According to the growth of national income, people nowadays pay much more attention to the health issues; Corresponsively the medical care gets improved as well. Today, we have advanced medical devices such as ultrasonic instrument, computed tomography, position emission tomography and magnetic resonance imaging, in which the ultrasonic instruments are now in widespread used because of low-cost, non-side effect, mobile and non-aggressive superiorities. Read more


Snake, also known as active contour model, is widely used in searching for region of interesting (ROI) in different images. However, Snake has many weak points, and they do deeply affect the range of application, especially in the use of medical images. Many researchers tried to modify Snake algorithm to increase the performance, however, they were hardly to be used widely, because they were only used in specific kinds of medical image. Read more


As digital videos become more and more popular nowadays, the application of videos in different fields has been wide spreading. The researches had turned from simply getting the shots to using techniques like abstraction, summarization to display the exciting performances in the clips; or even increasing audio techniques to let users listen to music while watching the films. Read more


The paper presents a new framework of video summarization which can extract and summarize video shots that a user interests in from a long and complicated video, according to their similarity of motion type and scene. Firstly, the shot detection adopts the color and edge information to make shot boundaries accurate. Read more


隨著數位化攝錄影設備的普及,不論是電視錄製的影片或是家用影片取得的管道也越來越便利,針對數位影片作為研究發展的領域也日益增長,從早期的Shot detection,進而發展到Video content analysis、Video structure analysis、Abstraction、High extraction等應用,使得影片可以更方便的觀賞及保存。電視上影片剪輯效果以往都需要經過專業的訓練,在這篇論文裡,我們提供了一個應用在多視角拍攝的棒(壘)球影片自動化剪輯系統,並且藉由推導出來的數學模型來模擬專業剪輯師的剪輯手法,讓個人化影片及家用影片也可以達到專業的編輯效果。 Read more


To fill the missing area seamlessly from the existing information of a damaged image is a difficult but interesting problem. Recently, many researchers pay their attention to find more effective algorithms. Image inpainting is such a useful technique which can be used to fill the miss area or damaged region automatically. In the thesis, we propose a robust algorithm to repair the missing region of an image, especially for large area. Read more


Inbetweening is very important in animation and movies. The thesis constructs an automatic system based on Inbetweening technique. After analyzing and drawing the curves, we convert two hand-drawn images into a smooth animation with colored shape. It is very sensitive after thinning. In order to get an accurate traveling order, we develop a strong tracing algorithm to solve simple image’s matching problem. Moreover, we use Bézier to approximate these curves, and use least square error estimation to find a control point. Finally, we have completed an animation. Read more


Digital painting synthesizes an output image with paint styles of example images along with the input source image. However, the synthesis procedure always requires the user intervention in selecting patches from example images that best describe its paint styles. The thesis presents a systematic system framework to synthesize example-based rendering images requires no user intervention in the synthesis procedure. Read more


The principal objective of this paper is to build a novel video tooning technology, and it contains two principal components: hierarchical video segmentation and video tooning with special effects. In the first part of the system, we through the human motion build a video structure, which includes Entry shot, Action shot, and Exit shot respectively. The Action shot is the core of the video, which is divided into Gimmickry period, Incubate period, and Shooting period. Read more


Matting and Compositing is originally used to deal with special effects composition problem in the movie technology, that is, in the boundary regions of object and background intersection, if it is a more complicated texture, such as tree, hair, water, smoke and some properties of light transmittance etc., the refraction of light will make pixel to receive the lights that come from object and background simultaneously, and this proportion of mix is called “Alpha Value”. Read more


As the more security demand increases, the intelligent monitor system has attracted considerable attention in recent years. In the previous related research, much effort is devoted to the detection of threat or dangerous situation. But the security application does not pay attention to the condition when the person is identified as no threat but needs assistance. Read more


Painting style transformation aims at producing a painting of one artist’s style from picture or other paintings. In the previous works of example-based painting style transfer, users have to select patches which describe painting style best from example patches manually. It is time-consuming and subjective for a user to select proper patches from a large number of patches. Read more


With the rise of Chinese worldwide, traditional Chinese art works have become much more valued than before. Many works of renowned painters have been steadily poured into the world’s major auctions, and that their values often hit a new record high. Nonetheless, when we take a look at the recent Chinese history, those frequent wars and invasions from superpower countries have made preservation of these paintings quite a challenge. Although many precious artistic works have been luckily survived from such a disturbing period; the marks left on them by age and wars, however, cannot be easily erased. Read more

建立以聲學特徵為基礎 應用於國語混淆發音之輔助學習系統

國語教學在台灣地區雖然自小學即開始,甚至在幼稚園就開始教注音符號, 惟在國語注音的教學上,一般老師仍是以口頭直接教學法發音,部分輔以拼音卡 等教具,先由老師示範,學生跟著唸讀,老師再以指正方式進行語音糾正。此傳 統方式若碰到教學者本身發音有瑕疵,或學生對發音部位及方法掌握不當,再加 上某些語音的特有的地域性差異,例如閩南語中沒有ㄓ、ㄔ、ㄕ的舌尖後音,也 沒有ㄈ的語音,因此容易將「發生」說成「花生」。這些因素將造成學習者對一 些易混淆的發音無法準確掌握及分辨,在語言表達上很容易讓人誤解,尤其是對 有心從事大眾媒體傳播業的新聞或廣播之人員而言,字正腔圓更形重要。本論文 主要研究動機,即希望透過收集從事廣播之播音員所錄製之國語注音發音及容易 發音混淆之詞句的發音,歸納分析出可用於判斷國語發音是否混淆的辨識方法, 以作為國語發音輔助學習之用。 Read more


人體動作追蹤在電腦視覺這塊領域已經發展一陣子了,至今還是相當熱門的研究題目,且應用的範圍相當多,如監控系統、家庭看護、虛擬實境以及遊戲。追蹤人體動作參數有兩種不同的方法,分成使用標記與不使用標記的追蹤。前者在人體身上攜帶標記透過計算標記的位置來模擬人體動作,例如:電影動畫是利用感應器(Motion Capture Sensor)裝置在人體的各個部位,根據回傳的參數得到人體三度空間的座標位置。後者是人體不使用標記做追蹤,其中又分成多台相機系統與單一相機系統,多台相機系統乃經由不同的角度取得相機間人體相對應的座標位置進行追蹤,或利用單一相機捨去深度的資訊作物體的追蹤。 Read more


社交網路分析一直以來都是個熱門的話題,所有團體成員彼此之間社交關係 的集合就是這個團體的社交網路,而透過社交網路分析可以瞭解團體成員之間的 互動,這分析可應用在各種與人有關的領域上。在學校裡,學生之間小團體的組 成及班級中領導人物與被孤立者的存在,一直都是教育者相當關心的部份。在團 體精神治療中,成員之間的交流情況是分析治療成果的指標之一。在網路社群中, 瞭解使用者群體之間的互動可以幫助廠商開發更人性化的網路產品。 Read more

使用主動式可疑區域萃取和增強式EM方法之 重製影像偽造偵測系統

在影像處理軟體和數位影像普及下,數位偽造影像偵測領域越來越受到重視。隨著影像資訊量越來越大和驗證演算法過程往往牽涉大量的數學運算之下,以往常見的人工框取ROI或是以sliding window做全域搜尋的方法顯得較不適用。前者仰賴大量人工的介入,後者則是消耗大量的時間成本。有鑑於此,我們提出一個重製偽造影像的自動偵測快篩系統,以減少人工介入和維持偵測精確度中間取得折衷為目標。 Read more


當我們使用數位攝影機進行影片的拍攝時,常會遇到動態範圍不足的問題,動態範圍指的是相機所能表現實際場景中最亮到最暗的亮度範圍,因實際拍攝時場景的亮度範圍時常遠高於數位攝影機的動態範圍,容易出現曝光過度的高光區域或曝光不足的陰影區域,無論是高光或陰影區域都有著亮度值平坦,缺乏細節的特性,因人眼的動態範圍較數位相機要大得多,可以清楚看到數位攝影機拍攝到的高光和陰影處之影像,因此數位攝影機與人眼所見的視覺效果有很大的差異,為了在觀賞上更為接近人眼所見,因此我們提出一個新穎的方法針對這些高光和陰影區域做修補,使得這些區域可以增加更多的細節,以達到更接近人眼所見之目的。 Read more


隨著攝影器材的發展越來越多人能自行拍攝私人影片,有些影片為了娛樂性和觀賞性加入動作情節並且利用特效來突顯動作情節,而一般這種特效大多是人工手動製作加入往往是花費大量時間而且是需要經驗的。 Read more


隨著近幾年人機介面(HCI)科技的發展,藉由身體來操作的互動系統相關的研究越來越受到人們的重視。舉例來說,像是一些多媒體的周邊儀器如Kinect、Wii remote controller跟PlayStation Moven所提供的直覺操控能力,使用者可以藉由擺動他們的雙手或身體來玩遊戲;另外有些行動裝置如iPhone、iPad所提供的觸碰螢幕界面,使用者可以利用手指的滑動或觸碰螢幕來進行操作。直覺且自然的低限制操作方式,在HCI領域裡儼然成為一種趨勢。 Read more

不完整數據下基於半監督式學習之 年齡估測

由於有廣泛的應用價值,近年來年齡估測的相關研究引起許多研究者的興趣。已有很多文獻提出年齡的特徵模型和估測模型,但是年齡估測尚有主要難以解決的問題。如年齡多樣性(Aging diversity)和年齡不充分數據(Insufficient aging data)使得估測真實年齡的精準度仍有限制,而本論文將目標放在年齡訓練數據不充分此問題上的處理。以實際面來看,年齡影像收集不易甚至在資料庫裡有些年齡樣本特別稀少,要透過人工標記也需要耗費大量的成本,因此在不足量的年齡標記樣本學習之下可能會限制住估測模型的學習推廣能力。 Read more

基於耦合隱藏式馬可夫模型之 桌球招式辨識

在桌球的領域中,人們依靠人眼觀察運動員的動作,進而分析其使用的招式與打法風格。然而桌球是一項技巧性相當高的運動,其中所使用的招式千變萬化,需要足夠的背景知識與經驗才能夠做出正確的判斷,這樣的情形使桌球運動的情報蒐集更具有難度,同時讓許多不熟悉桌球運動的觀眾難以分辨這些招式,進而造成觀眾無法體會一場精彩比賽的箇中奧妙。因此如何使用電腦視覺技術辨識桌球招式就變得十分的重要。 Read more