Appearance matching of occluded objects using coarse-to-fine adaptive masks

Edwards, Jaety; Murase, Hiroshi

doi:10.1109/cvpr.1997.609377

Cited by 20 publications

(14 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1 A search window, which is the ANDed area of the object regions of all images in the training image set, can also be used to deal with the background clutter [29], however, this does not help with occlusion. This search window technique can be extended to an adaptive mask to deal with the occlusion [30], however, occluding objects must be limited to a predefined set. Rather than using fixed masks, Leonardis et al [34,37] randomly select pixels from an image in order to identify the most likely object.…”

Section: Eigendecomposition Applied To Pose Estimationmentioning

confidence: 99%

“…Unfortunately, one of the drawbacks associated with using eigendecomposition-based approaches is that they are very sensitive to occlusion and background clutter [24,[28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44]. The purpose of this work is to explore the feasibility of applying eigendecomposition to a quadtree representation of correlated images in order to efficiently accommodate the presence of occlusion and background clutter.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Quadtree-based eigendecomposition for pose estimation in the presence of occlusion and background clutter

Chang

Maciejewski

Balakrishnan

et al. 2006

Pattern Anal Applic

View full text Add to dashboard Cite

Eigendecomposition-based techniques are popular for a number of computer vision problems, e.g., object and pose estimation, because they are purely appearance based and they require few on-line computations. Unfortunately, they also typically require an unobstructed view of the object whose pose is being detected. The presence of occlusion and background clutter precludes the use of the normalizations that are typically applied and significantly alters the appearance of the object under detection. This work presents an algorithm that is based on applying eigendecomposition to a quadtree representation of the image dataset used to describe the appearance of an object. This allows decisions concerning the pose of an object to be based on only those portions of the image in which the algorithm has determined that the object is not occluded. The accuracy and computational efficiency of the proposed approach is evaluated on 16 different objects with up to 50% of the object being occluded and on images of ships in a dockyard.

show abstract

Section: Eigendecomposition Applied To Pose Estimationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Quadtree-based eigendecomposition for pose estimation in the presence of occlusion and background clutter

Chang

Maciejewski

Balakrishnan

et al. 2006

Pattern Anal Applic

View full text Add to dashboard Cite

show abstract

“…(David & DeMenthon, 2005) proposed a method to use model and image line features to locate complex objects in high clutter environments. In appearance-based object recognition, (Edwards & Murase, 1997) addressed the occlusion problem inherent in appearance-based methods using a mask to block out part of the basic eigenimages and the input image. (Leonardis & Bischof, 1996) handled occlusion, scaling, and translation by randomly selecting image points from the scene and their corresponding points in the basis eigenvectors.…”

Section: Related Workmentioning

confidence: 99%

Recognition of Partially Occluded Elliptical Objects using Symmetry on Contour

Cho¹,

Choi²

2007

Vision Systems: Segmentation and Pattern Recognition

View full text Add to dashboard Cite

“…Edwards and Murase [28] used a correlation based approach in the spatial domain in order to detect multiple copies of one reference object. Smolińska [23] employed a scanning aperture for the correlation of small parts of a complex scene with simple patterns like circles.…”

Section: Introductionmentioning

confidence: 99%

Iterative nonlinear joint transform correlation for the detection of objects in cluttered scenes

Haist

Tiziani

1999

Optics Communications

View full text Add to dashboard Cite

Abstract-An iterative correlation technique with digital image processing in the feedback loop for the detection of small objects in cluttered scenes is proposed. A scanning aperture is combined with the method in order to improve the immunity against noise and clutter. Multiple reference objects or different views of one object are processed in parallel. We demonstrate the method by detecting a noisy and distorted face in a crowd with a nonlinear joint transform correlator.

show abstract

Appearance matching of occluded objects using coarse-to-fine adaptive masks

Cited by 20 publications

References 13 publications

Quadtree-based eigendecomposition for pose estimation in the presence of occlusion and background clutter

Quadtree-based eigendecomposition for pose estimation in the presence of occlusion and background clutter

Recognition of Partially Occluded Elliptical Objects using Symmetry on Contour

Iterative nonlinear joint transform correlation for the detection of objects in cluttered scenes

Contact Info

Product

Resources

About