Image processing apparatus, image processing method, and storage medium
US-2024428519-A1 · Dec 26, 2024 · US
US9530240B2 · US · B2
| Field | Value |
|---|---|
| Publication number | US-9530240-B2 |
| Application number | US-201314023168-A |
| Country | US |
| Kind code | B2 |
| Filing date | Sep 10, 2013 |
| Priority date | Sep 10, 2013 |
| Publication date | Dec 27, 2016 |
| Grant date | Dec 27, 2016 |
A practical reading order for non-experts. Skip the full description unless you need deep technical detail.
What the patent document calls the invention.
A short plain-language summary of the technical disclosure.
Who owns or filed the patent and who is credited as inventor.
Filing, priority, publication, and grant dates set the timeline.
The legal scope of protection — read this for what is actually claimed.
Technology tags used to group this patent with similar filings.
Prior art links and similar publications in this corpus.
Official abstract text for this publication.
A method including receiving a first image of a scene captured from a first perspective, the first image including an object and a background; segmenting the first image to extract a first two-dimensional contour of the object; approximating a plurality of three-dimensional locations of a plurality of points on the first contour; generating a three-dimensional billboard of the object based on the three-dimensional locations; and projecting the first image onto the three-dimensional billboard.
Opening claim text (preview).
What is claimed is: 1. A method, comprising: receiving a first image of a scene captured from a first perspective, the first image including an object and a background; segmenting the first image to extract a first two-dimensional contour of the object; receiving a second image of the scene captured from a second perspective, the second image including the object and the background; segmenting the second image to extract a second two-dimensional contour of the object; determining a plurality of correspondences, wherein each of the plurality of correspondences matches a first point from the first two-dimensional contour with a second point from the second two-dimensional contour; approximating a three-dimensional contour based on the plurality of correspondences; generating a three-dimensional billboard of the object based on interpolation of the three-dimensional contour; and projecting at least one of the first and second image onto the three-dimensional billboard. 2. The method of claim 1 , wherein the approximating comprises determining three-dimensional locations based on the first contour and a constraint of the first contour. 3. The method of claim 2 , wherein the constraint is a placement of a first point of the first contour on a known surface. 4. The method of claim 1 , further comprising: triangulating a best three-dimensional location for each of the plurality of correspondences to approximate the three-dimensional contour. 5. The method of claim 4 , wherein the triangulating comprises determining the best three-dimensional location as a closest point of a plurality of rays based on the first point of the correspondence and the second point of the correspondence. 6. The method of claim 1 , wherein determining the plurality of correspondences comprises using a dynamic time warping technique on the first two-dimensional contour and the second two-dimensional contour. 7. The method of claim 6 , wherein the dynamic time warping technique is a modified dynamic time warping technique. 8. The method of claim 6 , wherein the dynamic time warping technique minimizes a value of a global cost function. 9. The method of claim 8 , wherein the global cost function comprises a combination of a visual distance between two points and an epipolar distance between the two points. 10. The method of claim 1 , further comprising: approximating a plurality of three-dimensional locations of a plurality of points on the first contour, and filtering the plurality of three-dimensional locations to produce a filtered plurality of three-dimensional locations, wherein the three-dimensional billboard is generated based on the filtered plurality of three-dimensional locations. 11. The method of claim 1 , further comprising: inflating the three-dimensional billboard based on a distance function transform to produce an inflated three-dimensional billboard, wherein the first image is projected onto the inflated three-dimensional billboard. 12. A system, comprising: an image capture apparatus comprising a first camera, the first camera capturing a first image of a scene from a first perspective, the first image including an object and a background; and an analysis apparatus comprising a memory including a set of instructions and a processor executing the set of instructions to perform operations comprising: receiving the first image from the image capture apparatus; segmenting the first image to extract a first two-dimensional contour of the object; wherein the image capture apparatus further comprises a second camera, the second camera capturing a second image of the scene from a second perspective, the second image including the object and the background, and wherein the operations further comprise: receiving the second image of the scene; segmenting the second image to extract a second two-dimensional contour of the object; determining a plurality of correspondences, wherein each of the plurality of correspondences matches a first point from the first two-dimensional contour with a second point from the second two-dimensional contour; approximating a three-dimensional contour based on the plurality of correspondences; generating a three-dimensional billboard of the object based on an interpolation of the three-dimensional contour; and projecting at least one of the first and second image onto the three-dimensional billboard. 13. The system of claim 12 , wherein the approximating comprises determining three-dimensional locations based on the first contour and a constraint of the first contour. 14. The system of claim 13 , wherein the constraint is a placement of a first point of the first contour on a known surface. 15. The system of claim 12 , wherein the approximating the three-dimensional locations comprises: triangulating a best three-dimensional location for each of the plurality of correspondences to approximate the three-dimensional contour. 16. The system of claim 15 , wherein the triangulating comprises determining the best three-dimensional location as a closest point of a plurality of rays based on the first point of the correspondence and the second point of the correspondence. 17. The system of claim 12 , wherein determining the plurality of correspondences comprises using a dynamic time warping technique on the first two-dimensional contour and the second two-dimensional contour. 18. The system of claim 17 , wherein the dynamic time warping technique is a modified dynamic time warping technique. 19. The system of claim 17 , wherein the dynamic time warping technique minimizes a value of a global cost function. 20. The system of claim 19 , wherein the global cost function comprises a combination of a visual distance between two points and an epipolar distance between the two points. 21. The system of claim 12 , wherein the operations further comprise: approximating a plurality of three-dimensional locations of a plurality of points on the first contour, and filtering the plurality of three-dimensional locations to produce a filtered plurality of three-dimensional locations, wherein the three-dimensional billboard is generated based on the filtered plurality of three-dimensional locations. 22. The system of claim 12 , wherein the operations further comprise: inflating the three-dimensional billboard based on a distance function transform to produce an inflated three-dimensional billboard, wherein the first image is projected onto the inflated three-dimensional billboard. 23. A non-transitory computer-readable storage medium storing a set of instructions executable by a processor, the set of instructions, when executed by the processor, causing the processor to perform operations comprising: receiving a first image of a scene captured from a first perspective, the first image including an object and a background; segmenting the first image to extract a first two-dimensional contour of the object; receiving a second image of the scene captured from a second perspective, the second image including the object and the background; segmenting the second image to extract a second two-dimensional contour of the object; determining a plurality of correspondences, wherein each of the plurality of correspondences matches a first point from the first two-dimensional contour with a second point from the second two-dimensional contour; approximating a three-dimensional contour based on the plurality of correspondences; ge
Physics · mapped topic
Image-based rendering · CPC title
Edge-based segmentation · CPC title
Related publications grouped by family.
Answers are generated from the same data shown on this page.