P. Eisert, B. Girod. Model-based estimation of facial expression parameters from image sequences, in: Internat. Conf. on Image Processing, 1997, pp. 418--421.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....between two sets of faces of two different people performing similar facial expressions. In general it is hard to model and animate faces, even when they are cartoons characters. Usually complex models encoding the physical underlying musculature of the face are used (e.g. Candide model [1, 10]) Recently De la Torre [6] has used eigenfeatures [15, 17] to automatically learn person speci c appearance face models and dynamically animate them. The face is modeled using separate eigenfeatures since facial features such as the eyes and mouth undergo almost independent changes in appearance ....

P. Eisert and B. Girod. Model-based estimation of facial expression parameters from image sequences. ICIP, pp. 418-421, 1997.

No context found.

P. Eisert and B. Girod, "Model-based estimation of facial expression parameters from image sequences", International Conference on Image Processing, Oct. 1997.

No context found.

P. Eisert and B. Girod. Model-based estimation of facial expression parameters from image sequences. In Proceedings International Conference on Image Processing, 2:418--421, Oct. 1997.

....model and the motion constraints used by the facial expression synthesis are incorporated into the parameter estimation as described in the following. 1) Facial Parameter Estimation: In our model based coder all FAP s are estimated simultaneously using a hierarchical optical flow based method [21]. We employ a hierarchy of three spatial resolution layers with CIF as the highest resolution and each subsequent lower resolution layer subsampled by a factor of two, vertically and horizontally. We use the whole picture 348 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 10, ....

P. Eisert and B. Girod, "Model-based estimation of facial expression parameters from image sequences," in Proc. Int. Conf. Image Processing, vol. 2, Santa Barbara, Oct. 1997, pp. 418--421.

....3 Analysis of Facial Animation Parameters from Video Sequences While the MPEG 4 standard does not specify how to generate FAPs, we need a method which determines FAPs from video sequences in order to retrieve embedded watermarks from rendered sequences. A suitable method has been presented in [4, 5] and was used to obtain the FAPs from rendered sequences in the section on experimental results. The approach for the estimation of the facial parameters is model based and combines a motion model of an explicit 3D textured wireframe with the optical flow constraint from the video data. This ....

....with low computational complexity. The inputs to the estimation method are the video sequence and the 3D model of the head shown in the sequence, and the output are the estimated MPEG4 animation parameters. The method estimates the FAPs with very satisfying accuracy. For details, please refer to [4, 5]. 4 Watermarking of MPEG 4 Facial Animation Parameters For embedding of watermark data into the FAPs, we adopt a spread spectrum approach [6] that has been applied similarly to image and video watermarking before [3, 7, 8] The idea is to apply small changes to the FAPs that seem random and are ....

[Article contains additional citation context not shown here]

P. Eisert and B. Girod. Model-based estimation of facial expression parameters from image sequences. International Conference on Image Processing, 2:418--421, Oct. 1997.

....head model and the motion constraints used by the facial expression synthesis are incorporated into the parameter estimation as described in the following. C. 1 Facial Parameter Estimation In our model based coder all FAPs are estimated simultaneously using a hierarchical optical ow based method [21]. We employ a hierarchy of 3 spatial resolution layers with CIF as the highest resolution and each subsequent lower resolution layer subsampled by a factor of two vertically and horizontally. We use the whole picture of the face for the estimation, in contrast to feature based approaches, where ....

P. Eisert and B. Girod, \Model-based estimation of facial expression parameters from image sequences", in Proc. International Conference on Image Processing, Santa Barbara, Oct. 1997, vol. 2, pp. 418-421.

....head model and the motion constraints used by the facial expression synthesis are incorporated into the parameter estimation as described in the following. C. 1 Facial Parameter Estimation In our model based coder all FAPs are estimated simultaneously using a hierarchical optical flow based method [16]. We currently employ a hierarchy of 3 spatial resolution layers with CIF as the highest resolution and each subsequent lower resolution layer being subsampled by a factor of two vertically and horizontally. We use the whole picture of DRAFT, March 8, Eisert, Wiegand, Girod: ....

P. Eisert and B. Girod, "Model-based estimation of facial expression parameters from image sequences", International Conference on Image Processing, vol. 2, pp. 418--421, Oct. 1997.

....3 Analysis of Facial Animation Parameters from Video Sequences While the MPEG 4 standard does not specify how to generate FAPs, we need a method which determines FAPs from video sequences in order to retrieve embedded watermarks from rendered sequences. A suitable method has been presented in [14,15] and is outlined in the following. We assume that the human head is locally deformable and that facial expressions can be represented by a linear combination of small elementary local movements. These movements are described by MPEG 4 FAPs. We further assume that we have a 3D model of the person ....

....which shows the head model in relaxed state, i.e. FAP k ( Gamma1) 0 8 k (16) We then gain the absolute values of the FAPs for each frame by cumulative addition over DeltaFAP for all frames from I( 1) up to the current frame. The described method estimates the FAPs with very satisfying accuracy [15]. 4 Watermarking of MPEG 4 Facial Animation Parameters For embedding of watermark data into the FAPs, we adopt a spread spectrum approach [23] that has been applied similarly to image and video watermarking before [5,8,9] The idea is to apply small changes to the FAPs that seem random and are ....

P. Eisert and B. Girod. Model-based estimation of facial expression parameters from image sequences. In Proceedings International Conference on Image Processing, 2:418--421, Oct. 1997.

....positions. Figure 4: Camera frame of a training sequence (left) and extracted marker regions (right) 3. 2 Animation Parameters from 2D Correspondences For the estimation of facial animation parameter changes between two successive frames we use a modified version of the method proposed in [11]. Instead of using the optical flow constraint equation, the approach has been adapted for the use of 2D point correspondences. We can set up an explicit linear function for each marker that defines its displacement vector as a function of the unknown facial animation parameters (FAPs) This ....

P. Eisert and B. Girod, "Model-based estimation of facial expression parameters from image sequences", International Conference on Image Processing, Oct. 1997.

No context found.

P. Eisert, B. Girod. Model-based estimation of facial expression parameters from image sequences, in: Internat. Conf. on Image Processing, 1997, pp. 418--421.

No context found.

P. Eisert, B. Girod. Model-based estimation of facial expression parameters from image sequences, in: Internat. Conf. on Image Processing, 1997, pp. 418--421.

No context found.

P. Eisert and B. Girod, "Model-based estimation of facial expression parameters from image sequences," in Proc. of the IEEE Intl. Conference on Image Processing, October 1997, vol. 2, pp. 418--421.

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