G.J. Klinker, S.A. Shafer, and T. Kanade, "Image Segmentation and Reflection Analysis Through Color", Proc. SPIE #937: Applications of Artificial Intelligence VI, SPIE, Orlando, Florida, April 4-8 1988, pp. 229-244, Also in the proceedings of the DARPA Image Understanding Workshop, April 6-8, 1988, Boston, MA  Home/Search   Document Details and Download   Summary   Related Articles   Check

....of both reflection components, we can separate the two components by simply projecting the red, green, and blue irradi ance signals to the two vectors representing the reflection components in the (E, space. Separation of reflection components had been studied by by Klinker, Shafer, and Kanade [4], and by Gershon, Jepson, and Tsotsos [2] The former group searches for a skewed T signature, while the latter searches for a dog leg distribution in three dimensional color space. Both methods require that the specular reflection component rise and fall in a very short spatial distance ....

G.J. Klinker, S.A. Shafer, and T. Kanade, "Image segmentation and reflection analy- sis through color," Proceedings of Image Understanding Workshop, Cambridge, Massachusetts, April 1988, pp. 838-853.

....reflections can change the distribution of pixel colors between two views as shown in Figure 3. Recently there have been some algorithms for the detection of specularity based on physical models of light reflection. Those algorithms use structured lighting, color or polarization [7] 9] 11] [5] [4] 1] However most of the algo7 rithms require many restrictive assumptions on object reflectance and on illumination color and direction. The spectral differencing algorithm, on the other hand, does not require any assumptions on material type, reflectance variation or on illumination color ....

G.J. Klinker, S.A. Shafer, and T. Kanade. Image segmentation and reflection analysis through color. In Proceedings of the DARPA Image Understanding Workshop, pages 838--853, Pittsburgh, PA, 1988.

....segmentation nor the neighbourhood based segmentation has used underlying physical models of the colour image formation process in developing colour difference metrics. Physically based algorithms have produced good segmentation for colour images obtained under controlled conditions [22], 23] 24] 19] 3.4 Physically based segmentation Most colour image segmentation methods that are based on a very simple interpretation of colour changes in an image generally segment images not only along material boundaries but also along other lines exhibiting colour or intensity variations, ....

....colour changes in an image generally segment images not only along material boundaries but also along other lines exhibiting colour or intensity variations, such as highlight and shadow boundaries, or object edges with significant shading changes. To solve this problem, Klinker, Shafer and Kanade [22] [23] 24] present an approach to segment surfaces with colour variations due to highlights and shading. They establish a model called the Dichromatic Reflection Model. Materials can be divided into two classes on the basis of their optical properties, that is, optically homogeneous and optically ....

G. J. Klinker, S. A. Shafer, and T. Kanade, "Image segmentation and reflection analysis through colour," Proc. of Image Understanding Workshop, Cambridge, Massachusetts, April 1988, pp. 838-853.

.... has led to the use of deformable shape models in image segmentation [8, 10, 26, 28, 30, 38, 44, 45] Another strategy is to utilize image features that are somewhat invariant to illumination [7, 24] or to directly model the physics of illumination, color, shadows, and surface inter reflections [20, 23, 32, 33]. Such physicallybased approaches have also been shown to improve segmentation accuracy, and can be used to improve performance of model based methods. Unfortunately, the above mentioned techniques are going to make mistakes in merging regions, even in constrained contexts. This is because local ....

G.J. Klinker, S.A. Shafer, and T. Kanade. Image segmentation and reflection analysis through color. IJCV, 2(1):7--32, June 1988.

....that yield values which are invariant to scene illumination and viewpoint. To date, some investigation of II s has been conducted for visible imagery, but practically none has been reported for non visible imagery. Examples of II s in computer vision include color features for object recognition [14], 15] 16] and polarization cues for material identification [17] A more direct example of this approach computes ratios of albedos of homogeneous image intensity patches within objects in visible imagery [18] DRAFT May 17, ROBUST THERMOPHYSICS BASED INTERPRETATION 3 Non visible modalities ....

G.J. Klinker, S.A. Shafer and T. Kanade, "Image Segmentation and Reflection Analysis through Color", Proceedings of DARPA Image Understanding Workshop, Cambridge, MA, 1988, pp 838 - 853.

No context found.

G.J. Klinker, S.A. Shafer, and T. Kanade, "Image Segmentation and Reflection Analysis Through Color", Proc. SPIE #937: Applications of Artificial Intelligence VI, SPIE, Orlando, Florida, April 4-8 1988, pp. 229-244, Also in the proceedings of the DARPA Image Understanding Workshop, April 6-8, 1988, Boston, MA

No context found.

G. Klinker, S. Shafer, and T. Kanade. Image segmentation and reflection analysis through color. Int. J. Computer Vision, 2(1):7--32, June 1988.

No context found.

Gudrun J. Klinker, Steven A. Shafer, and Takeo Kanade. Image segmentation and reflection analysis through color. In Applications of Artificial Intelligence VI, volume 937, pages 229--244. SPIE, 1988.

No context found.

G.J. Klinker, S.A. Shafer and T. Kanade, "Image Segmentation and Reflection Analysis through Color", Proceedings of DARPA Image Understanding Workshop, Cambridge, MA, 1988, pp 838- 853.

No context found.

G.J. Klinker, S.A. Shafer and T. Kanade, "Image Segmentation and Reflection Analysis through Color", Proceedings of DARPA Image Understanding Workshop, Cambridge, MA, 1988, pp 838 - 853.

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