METHOD AND DEVICE FOR REAL-TIME MULTI-VIEW PRODUCTION

• US 20110025822A1
• Filed: 12/27/2007
• Published: 02/03/2011
• Est. Priority Date: 12/27/2007
• Status: Active Grant

• Current Assignee: Psholix AG
• Original Assignee: Sterrix Technologies UG
• Inventors: Naske, Rolf-Dieter
• Granted Patent: US 8,736,669 B2
• Time in Patent Office: Days
• US Class Current: 348/44
• CPC Class Codes: H04N 13/00 Stereoscopic video systems; Multi-view video systems; Details thereof
• CPC Class Codes: H04N 13/111 Transformation of image signals corresponding to virtual viewpoints, e.g. spatial image interpolation
• CPC Class Codes: H04N 13/20 Image signal generators
• CPC Class Codes: H04N 13/282 for generating image signals corresponding to three or more geometrical viewpoints, e.g. multi-view systems

Abstract

Methods and devices are disclosed, wherein from a stereo image having at least 2 perspectives additional (in general 8 or more) virtual perspectives can be produced, on condition that the time for the entire production process is less than the specified frame rate, and that simultaneously the multi-view image quality is maximized. It is not required that a disparity map exists. In general, the image regions that have not change between two stereo images of the image sequence are identified. In these regions, the associated disparity map also cannot have changed. In addition, the processing parameters are continuously adjusted through continuous time measurement during processing such that the quality is always at the maximum thereof. In this way, an adaptive method is obtained, which can also respond to load fluctuations of the underlying hardware device.

Claims

1. Method for the generation of m>

0 synthetic images from a stereo image of a stereo image sequence with two or more sub-images in such manner that a control unit measures the processing time and compares it with given target values and modifies control parameter in such manner as to render with maximum resolution and minimal image error, where the generation of the synthetic images being finished before the next stereo image in the sequence is received.

2. Method according to claim 1, wherein the resolution of the synthetic images can be modified by the user depending on the measured processing time for the synthetic images.

3. Method according to claim 1, wherein the resolution of the synthetic images can be modified by an automatic estimation method depending on the measured processing time for the synthetic images.

4. Method according to claim 1, wherein the number of used features for the disparity estimation can be modified by the user depending on the measured processing time for the synthetic images.

5. Method according to claim 1, wherein the number of used features for the disparity estimation can be modified by an automatic estimation method depending on the measured processing time for the synthetic images.

6. Method according to claim 1, wherein areas in a sub-image of a stereo image will be identified that deviate from the sub-image of the previous stereo image by less than a given value ε

, and wherein the computed disparity values of the corresponding pixels of the previous sub-image are re-used.

7. A device for the execution of a method according to claim 1 with a compressor for adjusting the received image size to the processing size, a constant detector for the identification of constant image areas, a feature extractor for the computation of features of the pixels, an image analyzer for the correspondence analysis, an image synthesizer for the generation of additional perspectives and a control unit that sets the control parameters, such that the computation of the additional perspectives is finished before a new stereo image is received.