Simultaneous Detection and Recognition of Dynamic Gestures by the A*-Guided Dynamic Time Warping and Progressive Shape Detection in Cluttered Scenes

View Statistics Email Alert RSS Feed

Information

Details

Project title

Code/計畫編號

NSC97-2221-E019-040

Translated Name/計畫中文名

使用A*導引式動態時間校準比對法與漸進式形狀偵測法在雜亂環境裡同時偵測與辨識動態姿勢

Project Coordinator/計畫主持人

Chin-Chun Chang

Funding Organization/主管機關

National Science and Technology Council

Department/Unit

Department of Computer Science and Engineering

Website

https://www.grb.gov.tw/search/planDetail?id=1684737

Year

2008

Start date/計畫起

01-08-2008

Expected Completion/計畫迄

31-07-2009

Bugetid/研究經費

400千元

ResearchField/研究領域

資訊工程--硬體工程

"動態姿勢辨識是發展先進人和電腦互動系統(human-computer interaction)的核心技術之一。這種系統的一個典型應用可為，根據事先定義的動作來辨識人的動作指令的命令解釋器。為了在一般環境下操作此系統，這種系統必須有能力在複雜環境中鎖定與分析目標肢體的動作。然而，這項技術目前仍舊是一個十分有挑戰性的問題。由於事先學習的動態動作樣版提供豐富的資訊來輔助鎖定目標肢體，我們擬將在這個三年計畫發展能在複雜環境裡同時偵測與辨識動態動作的技術。我們將整合A*搜尋(the A* search)，動態時間校準比對(dynamic time warping)，及漸進式形狀偵測法來在複雜環境裡同時偵測與辨識動態動作。在這裡A*搜尋將用來掌控整個偵測與辨識的程序以避免不必要的形狀偵測計算。動態時間校準比對將根據漸進式形狀比對的結果，逐步地並越來越精準地比對輸入影像序列與事先學習的動態姿勢樣本。在這個架構下，事先學習的動態姿勢樣本所含有的空間與時間上的資訊將會被自然的應用出來。這三年擬將研究的主題如下。 􀂄 第一年將整合 A*搜尋，動態時間校準比對，與一般霍夫轉換(generalized Hough transform)在一般複雜環境，來同時偵測與辨識動態動作。 􀂄 第二年將整合 A*搜尋，動態時間校準比對，與核技術為基礎的可變形形狀模型(kernel-based deformable shape models)在一般複雜環境，來同時偵測與辨識動態動作。 􀂄 第三年將研究如何藉由調適一個通用的動態動作模型來製作個人動態動作模型一般來減少學習動態動作樣本時間。同時我們也將研究是否可將前兩年的結果用部分結構(part-based)為主的方式來製作以容忍目標肢體更大的變形量。由於我們擬將研究的技術是建構在幾個已經發展成熟的技術上，如動態時間校準比對，A*搜尋，一般霍夫轉換，相關性過濾器(correlation filter)，核技術等，其結果十分值得期待。""Dynamic gesture recognition is one of the core technologies to develop advanced input devices for natural human-computer interaction. A typical application of such devices is a command interpreter, which classifies the activity of a performer into one of the pre-defined actions according to the pre-learned prototypical dynamic gesture patterns. In order to operate such devices in general environments, these devices should be capable of locating and analyzing the target body part in cluttered scenes, which is still a challenging problem. Since the pre-learned prototypical gesture patterns provide valuable cues for locating the target body part, we shall investigate the techniques of simultaneous detection and recognition of dynamic gestures in cluttered scenes in this project. In this three-year project, we shall integrate the A* search, the dynamic time warping algorithm, and the technique of progressive shape detection together to simultaneously detect and recognize dynamic gestures in cluttered scenes. Here, the A* search will be used to control the whole detection and recognition process to avoid unnecessary computation on shape detection. The dynamic time warping algorithm will be used to match the input image sequence and the prototypical dynamic gesture patterns based on progressively refined shape detection results. Based on the framework which will be developed, the space and time information embedded in the prototypical dynamic gesture patterns will be utilized naturally. The topics of the three-year project are as follows. 􀂄 In the first year, we shall aim at integrating three algorithms, namely, the generalized Hough transform, the dynamic time warping algorithm, and the A* search, together to simultaneously locate the body part and recognize the type of the action in a moderately cluttered scene. 􀂄 In the second year, we shall focus on integrating three algorithms, namely, the kernel-based deformable shape model, the dynamic time warping algorithm, and the A* search, together to simultaneously locate the target body part and recognize the type of the action in a moderately cluttered scene. 􀂄 In the third year, we shall focus on adaptation of generic dynamic gesture models of the target body part to those for an individual to reduce the complexity of the learning phase. Meanwhile, we shall study the possibility of a part-based implementation of the systems developed in the first two years to make the systems tolerate more shape deformation of the target body part. Since the technology which will be developed is based on several well developed algorithms such as the dynamic time warping algorithm, the A* search, the generalized Hough transform, the correlation filter, and the kernel method, the result of this project could be expected."

Keyword(s)

動態姿勢辨識
運動偵測
動態時間校準比對法
漸進式形狀偵測法
dynamic gesture recognition
human motion detection
dynamic time warping
progressive shape detection

DSpace CRIS

Simultaneous Detection and Recognition of Dynamic Gestures by the A*-Guided Dynamic Time Warping and Progressive Shape Detection in Cluttered Scenes

Details

Description