Coarse-to-fine adaptive masks for appearance matching of occluded scenes

Edwards, Jaety; Murase, Junichi

doi:10.1007/s001380050075

Cited by 11 publications

(10 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In terms of the immediate future of recognition systems, representation modules in computer vision will probably draw on the knowledge that is contained in the computer graphics field, possibly utilizing subdivision and multiresolution schemes which have emerged as major topics in that area. In addition, it is obvious that there is a glaring lack of methods in the current literature which have the ability to represent and [19] Splines Y Y 9 100% SAI [31] Angles Y ---Splash [59] Surface normals Y Y 9 -Point signatures [18] Distance Y Y 15 100% Spin images [33] Surface histogram Y Y 4 100% Greenspan [26] Voxels Y Y 5 -COSMOS [20] SSF N Y 30 97% Crease angle [6] Crease angles Y N 4 92% [21] Occlusion masks Y Y 6 89% Leonardis [41] Robust est. Y Y 20 75% Ohba [50] Eigenwindows Y N --OLIVER [54] Probabilistic models Y ---Mokhtarian [46] CSS Y Y 15 100% Vijayakumar [66] Silhouettes Y Y --Chen [14] Internal edges Y Y 60 -Sullivan [62] Errors N -2 100%…”

Section: Resultsmentioning

confidence: 99%

Representation and Recognition of 3D Free-Form Objects

Mamic

Bennamoun

2002

Digital Signal Processing

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 99%

Representation and Recognition of 3D Free-Form Objects

Mamic

Bennamoun

2002

Digital Signal Processing

View full text Add to dashboard Cite

“…. , u p ] (usually only p, p < n, eigenvectors are sufficient), an unknown sample x ∈ IR m can be reconstructed bỹ [17,18,19,20]). But since these methods are computationally very expensive (i.e., they are based on iterative algorithms) or can handle only a small amount of noise, they are often not applicable in practice.…”

Section: Fast-robust Pcamentioning

confidence: 99%

“…Due to its least squares formulation, PCA is highly sensitive to outliers. Thus, several methods for robustly learning PCA subspaces (e.g., [12,13,14,15,16]) as well as for robustly estimating the PCA coefficients (e.g., [17,18,19,20]) have been proposed. In this paper, we are focusing on the latter case.…”

Section: Introductionmentioning

confidence: 99%

Fast-Robust PCA

et al. 2009

View full text Add to dashboard Cite

Abstract. Principal Component Analysis (PCA) is a powerful and widely used tool in Computer Vision and is applied, e.g., for dimensionality reduction. But as a drawback, it is not robust to outliers. Hence, if the input data is corrupted, an arbitrarily wrong representation is obtained. To overcome this problem, various methods have been proposed to robustly estimate the PCA coefficients, but these methods are computationally too expensive for practical applications. Thus, in this paper we propose a novel fast and robust PCA (FR-PCA), which drastically reduces the computational effort. Moreover, more accurate representations are obtained. In particular, we propose a two-stage outlier detection procedure, where in the first stage outliers are detected by analyzing a large number of smaller subspaces. In the second stage, remaining outliers are detected by a robust least-square fitting. To show these benefits, in the experiments we evaluate the FR-PCA method for the task of robust image reconstruction on the publicly available ALOI database. The results clearly show that our approach outperforms existing methods in terms of accuracy and speed when processing corrupted data.

show abstract

“…Edwards and Murase [4] introduced a modification of the PCA technique, where robustness to partial occlusion was achieved by excluding all image pixels from the subspace projection that constituted occluded image areas. They developed an (coarse-to-fine) adaptive masking concept that enabled a robust calculation of the subspace representation.…”

Section: Introductionmentioning

confidence: 99%

“…To a certain extent our approach bears similarities with the approach of Edwards and Murase [4]; however, instead of simply excluding image pixels that constitute occluded image areas (i.e., masking the image with a binary mask), it weights each image pixel according to a confidence measure that quantifies the confidence that the pixel in question represents an outlier. By doing so, it reduces (or eliminates) the impact of the occluded pixels on the subspace representation of the given face image and adds robustness to partial occlusions of the underlying subspace projection technique.…”

Section: Introductionmentioning

confidence: 99%