M. Black, D. Fleet, and Y. Yacoob. Robustly estimating changes in image appearance. In Computer Vision and Image Understanding, volume 78, pages 8--31, Apr. 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....for Lambertian surfaces and with only ambient light present, which is far from realistic. A simple relaxation is to allow global brightness changes by adding a constant a to all points gray values. Further approximation is to include linear brightness changes in the image plane over the object. [5]. This crude model can be written as: M i(i) Mo(i) a [ b c ] a7im,i (9) where we have a dot product of the vector [ b c ] and image projection of the i th object point, vector im,i. The parameters a, b, c should be estimated for each new image k. 20 15 Figure 2. Function (I) for = ....

M. Black, D. J. Fleet, and Y. Yacoob. Robustly estimating changes in image appearance. Cornput. Vis. Image Understanding, 78(1):8-31, 2000.

....model discontinuities and to keep the representation of distant scene points separate. In the context of motion estimation, single view methods have relied on known models of 3D shape [2] or motion [1] to make 3D motion estimation a well posed problem, or have focused on improving the robustness [18, 19] and physical validity [20] of 2D motion estimation. Unfortunately, the use of known models limits the types of scenes that can be reconstructed, while the ill posed nature of single view 3D motion estimation makes it difficult to account for image variations in a way that is consistent with a ....

M. J. Black, D. J. Fleet, and Y. Yacoob, "Robustly estimating changes in image appearance," CVIU, v. 78, n. 1, pp. 8--31, 2000.

....images. One way to cope with the above difficulty would consist of morphing all testing images to equal (in shape) the learning one [64] In order to achieve this, we can compute the optical flow between the learning and testing images and then use this motion information to morph the test face [5, 6] to equal in shape the one used for training. Unfortunately, this cannot always be achieved; e.g. the eye area of a screaming image (Fig. 3(d) cannot be morphed to a neutral eyes expression, because in most of the cases the texture of the inside of the eyes is not available and cannot be ....

M.J. Black, D.J. Fleet, Y. Yacoob, "Robustly estimating changes in image appearance," Computer Vision and Image Understanding 78(1):8-31, 2000.

....approach exploit motion discontinuities. In [5] Black and Fleet developed a powerful Bayesian framework for detection and tracking of motion discontinuities. A translation and an occlusion parametric model are explicitly stipulated to represent and recognize local image motions. In another paper [6], Black et al. developed a generalized generative model to reflect multiple changes such as a#ne motion, illumination, specular reflection, and iconic changes. A robust statistical framework was used for recovering these appearance changes in image sequences. The last approach to be discussed ....

M.J. Black, D.J. Fleet, and Y. Yacoob. Robustly estimating changes in image appearance. Computer Vision and Image Understanding, 78(1):8--31, 2000.

....separating the external causes from the intrinsic properties in the appearance of each object. This task, roughly equivalent to perceptual constancies in human visual perception, is currently an active research area in computer vision. Several papers in the special issue address this topic [1, 3, 6, 8]. Robust statistical methods were first adopted in computer vision to improve the performance of feature extraction algorithms at the bottom level of the vision hierarchy. These methods tolerate (at various degrees) the presence of data points not obeying the assumed model. Such points are ....

....Image Understanding offer a representative sample for the state of the art approaches in robust computer vision. Several common trends can be recognized in these papers. The concept of robust analysis is extended to semantically more meaningful problems (higher levels of the visual hierarchy) in [1, 6], or to more realistic operating conditions in [5] A second trend is to combine robust estimators with other techniques to obtain algorithms with increased autonomy and better performance, overcoming some of the shortcomings outlined above. The combinations include: the maximum likelihood ....

[Article contains additional citation context not shown here]

M.J. Black, D.J. Fleet and Y. Yacoob. Robustly estimating changes in image appear8 ance. In this issue.

....separating the external causes from the intrinsic properties in the appearance of each object. This task, roughly equivalent to perceptual constancies in human visual perception, is currently an active research area in computer vision. Several papers in this special issue address this topic [1, 3, 6, 8]. Robust statistical methods were first adopted in computer vision to improve the performance of feature extraction algorithms at the bottom level of the vision hierarchy. These methods tolerate (in various degrees) the presence of data points that do not obey the assumed model. Such points are ....

....Understanding offer a representative sample of INTRODUCTION 3 the state of the art in robust computer vision. Several common trends can be recognized in these papers. The concept of robust analysis is extended to semantically more meaningful problems (higher levels of the visual hierarchy) in [1, 6] and to more realistic operating conditions in [5] A second trend is to combine robust estimators with other techniques to obtain algorithms with increased autonomy and better performance, overcoming some of the shortcomings outlined above. The combinations include the maximum likelihood paradigm ....

[Article contains additional citation context not shown here]

M. J. Black, D. J. Fleet, and Y. Yacoob, Robustly estimating changes in image appearance, Comput. Vision Image Understand. 78, 2000, 8--31.

.... still slightly curves especially around the corners of the image, because the gradation of illumination of the image can not be removed by the simple histogram transformation which should be replaced with an estimation of illumination change by some method, such as a linear brightness constraint[8]. Note that in simulation experiments using transformed pattern as a distorted image with additional noise at each pixel, the proposed method worked very well even when the amplitude of added uniform noise is greater than 50. 8. CONCLUSIONS We have proposed a new technique of automated camera ....

M. J. Black, D. J. Fleet, and Y. Yacoob, "Robustly estimating changes in image appearance, " CVIU, vol. 78, no. 1, pp. 8--31, 2000. (a) (b) (c) (d) (e) (f) (g) (h)

....accuracy. We illustrate the use of the 2D PSFAM model with several applications including video conferencing, realistic avatar animation and eye tracking. 1 Introduction Many computer vision researchers have used Principal Component Analysis (PCA) to parameterize appearance, shape or motion [3,10,23,30]. However, one major drawback of this traditional technique is that it needs normalized samples in the training data. In the case of computer vision applications, the result is that the samples have to be aligned or geometrically normalized (we assume that other normalizations, e.g. photometric, ....

....of PSFAMs are proposed. 3 Generative Face Models: Motivation Our eigen registration algorithm will be introduced with examples from face modeling. In this section we describe one possible generative model for dynamic faces. Similar to the previous work of Black and Jepson [4] Black et al. [3] and Cootes et al. 10] the generative model that we propose for image formation takes into account motion and appearance, but in our case we also exploit predefined masks and learn the appearance bases. Figure 3 shows some frames of a training set for learning a 2D PSFAM. Given this training ....

M. J. Black, D. J. Fleet, and Y. Yacoob. Robustly estimating changes in image appearance. Computer Vision and Image Understanding, 78(1):8--31, 2000.

....With only two frames, one can only model brightness changes that are linear in time. 2 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Hilton Head Island, SC, June 2000 If brightness is not conserved, then the optical flow field estimated from (2) can be severely biased [4, 17, 18, 24, 3, 20, 9, 19]. Causes of brightness variation include moving illumination envelopes, changing orientation of surfaces under directional illumination, and atmospheric influences in outdoor applications. Other instances occur in scientific applications that quantitatively investigate dynamic processes [13] ....

....estimates, and it allows us to estimate additional information that characterizes the physical processes. TLS error covariance matrices [21] are used to quantify the accuracy of the optical flow and the brightness change parameters. 2. Previous Work Brightness variations have been modeled by [4, 17, 18, 24, 3, 20, 9]. A general framework is proposed in [20] where the brightness change between two frames consists of a multiplier and an offset field: ### # ## # #### # ############## (3) where, for notational convenience, # # # # # ## # # denotes a space time 3D vector. It is certainly true that all ....

[Article contains additional citation context not shown here]

M. J. Black, D. J. Fleet, and Y. Yacoob. Robustly estimating changes in image appearance. CVIU 78:8--31, 2000.

No context found.

M. Black, D. Fleet, and Y. Yacoob. Robustly estimating changes in image appearance. In Computer Vision and Image Understanding, volume 78, pages 8--31, Apr. 2000.

No context found.

M. J. Black, D. J. Fleet, and Y. Yacoob. Robustly estimating changes in image appearance. CVIU, 78(1):8--31, 2000.

No context found.

M. J. Black, D.J. Fleet and Y. Yakoob, 2000 "Robustly Estimating Changes in Image Appearance", Computer Vision and Image Understanding. 78, p. 8-31.

No context found.

M.J. Black, D.J. Fleet, Y. Yacoob, " Robustly estimating changes in image appearance," CVIU, 78, 8--31, 2000.

No context found.

M. J. Black, D. J. Fleet, and Y. Yacoob, "Robustly estimating changes in image appearance," Comput. Vis. Image Understanding, vol. 78, pp. 8--31, 2000.

No context found.

M. Black, D. Fleet, and Y. Yacoob, "Robustly estimating changes in image appearance", Computer Vision and Image Understanding, 78(1):8--31, 2000.

No context found.

M. J. Black, D. J. Fleet, and Y. Yacoob, "Robustly estimating changes in image appearance," Computer Vision and Image Understanding, vol. 78, no. 1, pp. 8--31, 2000.

No context found.

Michael J. Black, David J. Fleet, and Yaser Yacoob. Robustly estimating changes in image appearance. Computer Vision and Image Understanding, 78, 2000.

No context found.

M. J. Black, D. J. Fleet and Y. Yacoob, "Robustly estimating Changes in Image Appearance", Computer Vision and Image Understanding, Vol. 78, No. 1, 2000.

No context found.

M. Black, D. Fleet, and Y. Yacoob, "Robustly Estimating Changes in Image Appearance", Computer Vision and Image Understanding, 78(1):8--31, 2000.

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