M. Yang and N. Ahuja. Recognizing hand gesture using motion trajectories. In CVPR, volume 1, pages 466--472, 1999. (b) (a) (b) (a)  Home/Search   Document Details and Download   Summary   Related Articles   Check

....images with per image using recursive partition trees 93.1 accuracy Neural Networks Decision available after Can not process a Schlenzig every frame (real time) sequence of frames et al. 126] Time Delay Neural Learns space time Need lots of Yang Networks patterns; shift invar. training data [149] Motion History Images Compact ecient Sensitive to viewing Bobick Motion Energy Images templates for perspective; does not [22] recognition handle spatial variation Dynamic Time Warping Time normalization of Behaves poorly with Gavrila (DTW) Continuous Dynamic space trajectories easy ....

M.H. Yang and N. Ahuja, \Recognizing hand gesture using motion trajectories, " IEEE Computer Society Conference on Computer Vision and Pattern Recognition,Fort Collins, CO, Vol. 1, pp. 466-72, June 23-25, 1999.

....edges. Human models have been initialized by hand in the first frame of each sequence [14] 18] Some authors have considered the problem of action activity gesture recognition for humans using shape and or motion information [8] 12] 20] 25] 26] 32] 36] 44] 45] 71] 82] 86] [87]. Dynamic recognition, most appropriate for interpreting video sequences, is done using recursive and neural networks, deformable templates, spatio temporal templates [60] and graphical models [12] because they offer dynamic time warping and a clear Bayesian semantics for both individual (HMM) ....

....examples include [18] 48] 72] 74] and [83] More recently, Bayesian inference has been used [22] 66] 73] 79] these methods are also known as Condensation. Other strategies that have been used include search algorithms such as best first [29] and or winner take all [18] 57] [87]. Bobick [11] recently proposed a taxonomy of movement, activity, and action. In his taxonomy, movements are primitives, requiring no contextual or sequence knowledge in order to be recognized. Activities are sequences of movements or states, where the only knowledge required to recognize them ....

M.-H. Yang and N. Ahuja, "Recognizing hand gestures using motion trajectories," in Proc. Computer Vision and Pattern Recognition, vol. 1, 1999, pp. 466--472.

....distribution over the interval. Gesture recognition is performed using a probabilistic finite state (event) machine. State transitions depend on both the observed model likelihood and the estimated state duration p.d.f. HMMs have been used extensively in visual gesture recognition recently [5,6,7,8,9]. HMMs are trained on data that is temporally well aligned. Given the sample data of a gesture trajectory, HMMs use dynamic programming to output the probability of the observation sequence. The maximum probability is compared with a threshold to decide if a gesture is recognized. In the ....

M.H. Yang and N. Ahuja, Recognizing Hand Gesture Using Motion Trajectories. CVPR99.

....were modeled as a number of blobs resulting from pixel classi cation based on their color and position in the image. Wren et al. 2] achieved segmentation by classifying pixels into one of several models, including a static world and a dynamic user represented by 2 gaussian blobs. Yang and Ahuja [3] used skin color and the geometry of palm and face regions for segmentation stages of their system. A Gaussian mixture (with parameters estimated by an EM algorithm) modeled the distribution of skin color pixels. Rehg and Kanade [4] used a 3 D hand model to track a hand. They compared line ....

M.-H. Yang and N. Ahuja. Recognizing hand gestures using motion trajectories. In Proc. of IEEE CS Conference on Computer Vision and Pattern Recognition, volume 1, pages 466-472, Fort Collins, CO, June 1999.

....data, using motion or appearance data. A majority of the methods began by segmenting the human body from the background. Wren et al. 2] achieved segmentation by classifying pixels into one of several models, including a static world and a dynamic user represented by gaussian blobs. Yang and Ahuja [3] used skin color and the geometry of palm and face regions for segmentation stages of their system. Yacoob and Black proposed parameterized representation of human movement [4] Cutler and Davis [5] segmented the motion and computed a moving objects self similarity (including human motion ....

M.-H. Yang and N. Ahuja. Recognizing hand gestures using motion trajectories. In Proceedings of IEEE CS Conference on Computer Vision and Pattern Recognition, volume 1, pages 466-472, Fort Collins, CO, June 1999.

....manual initialization, and can easily recover from occasional errors. 2 Related Work Several existing systems include automated hand detection and tracking. Such systems typically make restrictive assumptions on the domain: only hands move, the hands are the fastest moving objects in the scene [20, 37, 39, 7, 24, 26], hands are skin colored, or they are the only skincolored objects in the scene [24, 33] Often the background is assumed to be static, and known [24, 12] Some systems use such assumptions to obtain several possible regions This work was supported in part through Office of Naval Research Young ....

....Investigator Award N00014 96 1 0661, and National Science Foundation grants IIS 9912573 and EIA 9809340. Figure 1: Hand pose estimation overview. where the hands are, and use matching with appearancebased models to choose among those regions [37, 13] Stochastic tools, such as Kalman filtering [33, 37, 39], can be used to predict the hand position in a future frame. Overall, hand detection and tracking algorithms tend to perform well in restricted environments, where assumptions about the number, location, appearance and motion of hands are valid, and the background is known. Reliable performance ....

M. Yang and N. Ahuja. Recognizing hand gesture using motion trajectories. In CVPR, volume 1, pages 466--472,

No context found.

M. Yang and N. Ahuja. Recognizing hand gesture using motion trajectories. In CVPR, volume 1, pages 466--472, 1999. (b) (a) (b) (a)

No context found.

M.H. Yang and N. Ahuja, "Recognizing Hand Gesture Using Motion Trajectories," Proc. CVPR, I:1466-1472,1999.

No context found.

M.H. Yang and N. Ahuja, "Recognizing Hand Gesture Using Motion Trajectories," Proc. CVPR, I:1466-1472,1999.

No context found.

M.H. Yang and N. Ahuja, "Recognizing Hand Gesture Using Motion Trajectories," Proc. CVPR, I:1466-1472,1999.

No context found.

M. Yang and N. Ahuja. Recognizing hand gesture using motion trajectories. In CVPR, volume 1, pages 466--472,

No context found.

M. Yang and N. Ahuja. Recognizing hand gesture using motion trajectories. In Proceedings Computer Vision and Pattern Recognition, volume 1, pages 466-- 472, 1999.

No context found.

M.-H. Yang and N. Ahuja. Recognizing hand gestures using motion trajectories. In Proceedings of CVPR, volume 1, pages 466--472, Fort Collins, CO, June 1999. 6

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC