METHOD FOR PROCESSING A VIDEO DATA SET

• US 20100158482A1
• Filed: 05/02/2008
• Published: 06/24/2010
• Est. Priority Date: 05/04/2007
• Status: Active Grant

• Current Assignee: Imcube Labs GmbH
• Original Assignee: Imcube Media GmbH
• Inventors: Knorr, Sebastian, Sikora, Thomas
• Granted Patent: US 8,577,202 B2
• Time in Patent Office: Days
• US Class Current: 386/241
• CPC Class Codes: G06T 15/205 Image-based rendering
• CPC Class Codes: H04N 13/261 with monoscopic-to-stereoscopic image conversion
• CPC Class Codes: H04N 19/59 involving spatial sub-sampling or interpolation, e.g. alteration of picture size or resolution
• CPC Class Codes: H04N 19/597 specially adapted for multi-view video sequence encoding

Abstract

The invention relates to a method for processing a video data set, which comprises a sequence of original images, into a modified video data set which, in addition to the sequence of original images, comprises virtual images derived therefrom in one or more data processing units, wherein an SfM analysis is carried out for each of the original images and a respectively assigned original position of a recording device which was used to record the original image, virtual positions are formed for a recording device for recording virtual images, for each virtual position of the recording device with the associated virtual initial image, a respective assignment is created to at least one of the original positions of the recording device with the associated original image, for each assignment, a homography is determined between the virtual initial image and the associated original image, and for each virtual initial image, a virtual final image is formed from the associated original image in that the original image is formed in the associated final image using the respectively determined homography.

Claims

1. A method for processing a video data set, which comprises a sequence of original images, into a modified video data set which, in addition to the sequence of original images, comprises virtual images derived therefrom in one or more data processing units, wherein:

an SfM analysis is carried out for each of the original images and a respectively assigned original position of a recording device which was used to record the original image,

virtual positions are formed for a recording device for recording virtual images,

for each virtual position of the recording device with the associated virtual initial image, a respective assignment is created to at least one of the original positions of the recording device with the associated original image,

for each assignment, a homography is determined between the virtual initial image and the associated original image, and

for each virtual initial image, a virtual final image is formed from the associated original image in that the original image is formed into the associated final image using the respectively determined homography.

2. The method according to claim 1, characterized in that the assignments are formed in that at least one laterally closest original position of the recording device with the associated original image is assigned to each virtual position of the recording device with the associated virtual initial image.

3. The method according to claim 2, characterized in that the at least one laterally closest original position of the recording device is formed at a lateral distance of approximately 64 mm from the assigned virtual position of the recording device.

4. The method according to, claim 1 characterized in that the virtual positions for the recording device for recording the virtual images are formed at essentially equidistant spacings from the original positions.

5. The method according to, claim 1, characterized in that for each of the original positions, a plurality of virtual positions of the recording device is formed which are optionally equidistant from one another.

6. The method according to, claim 1, characterized in that the modified video data set is converted into a modified video data set with images of increased resolution in that, using the respectively determined homography, from the sequence of original images and the additional virtual images, a sequence of images with increased resolution is generated, said resolution being higher than an original resolution of the original images and the additional virtual images.