Full metadata record
DC FieldValueLanguage
dc.contributor.author顏宏年en_US
dc.contributor.authorYen, Hong- Nienen_US
dc.contributor.author張志永en_US
dc.contributor.authorChang, Jyh-Yeongen_US
dc.date.accessioned2015-11-26T01:06:56Z-
dc.date.available2015-11-26T01:06:56Z-
dc.date.issued2013en_US
dc.identifier.urihttp://140.113.39.130/cdrfb3/record/nctu/#GT070060064en_US
dc.identifier.urihttp://hdl.handle.net/11536/72519-
dc.description.abstract本篇論文實現一套自動化日夜居家監視系統,此系統為了提供良好的監控服務,著重於動作辨識和步態辨識,藉由步態辨識技術,掌控環境內每位成員的身份,其辨識動作以了解每個人的行動。本篇論文使用兩台攝影機在實驗室進行人物辨識及動作辨識。 動作辨識與步態辨識主要的資訊來自於人,擷取出人體部份為辨識的依據,為了更精確的擷取前景,使用灰階域與HSV色彩空間,建立兩種背景模型,並能有效的消除影像中陰影部分,使得擷取的前景能夠完整。接著將前景經由特徵空間轉換及標準空間轉換後,投影到維度較小的空間且能保有原影像的資訊。接著進行訓練,本方法加入時間資訊,將前景 5:1 減低抽樣取出影像,累積三張影像,建立模糊法則。辨識工作方面,使用預先學習且建立的模糊法則,進行辨識。  zh_TW
dc.description.abstractIn this thesis, we implement an automatic home health care system that combines action recognition and gait recognition in the day and night environments (bright and dark). Gait recognition can identify each person in the lab; action recognition can identify each person's actions. We use two cameras to recognize actions and gait, respectively. We build two background models, one in grayscale, and the other in the HSV color space, that extract the human region correctly. We also reduce the shadowing effect. For better efficiency, the binary image is transformed into a new space by eigenspace and canonical space transformation. Then we gathered three image frame sequence, 5:1 down sampling from the video, to convert to a posture sequence by template matching. The posture sequence is classified to an action or a person’s gait by fuzzy rules inference, which combines temporal sequence information for recognition.en_US
dc.language.isoen_USen_US
dc.subject動作辨識zh_TW
dc.subject步態辨識zh_TW
dc.subjectAction Recognitionen_US
dc.subjectGait Recognitionen_US
dc.title以模糊規則為基礎之日夜動作辨識及步態辨識zh_TW
dc.titleFuzzy Rule Based Day-and-Night Action Recognition and Gait Recognitionen_US
dc.typeThesisen_US
dc.contributor.department電控工程研究所zh_TW
Appears in Collections:Thesis


Files in This Item:

  1. 006401.pdf

If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.