Compression of stereo image pairs and streams
read more
Citations
Supporting Video-mediated Communication over the Internet
DSIC: Deep Stereo Image Compression
Stereoscopic video transmission over the Internet
Color stereoscopic images requiring only one color image
Scalable multi-view stereo camera array for real world real-time image capture and three-dimensional displays
References
Wavelets and signal processing
Data compression of stereopairs
Interpolative multiresolution coding of advance television with compatible subchannels
On stereo image coding
Constrained disparity and motion estimators for 3DTV image sequence coding
Related Papers (5)
Frequently Asked Questions (16)
Q2. What are the future works mentioned in the paper "Compression of stereo image pairs and streams" ?
Future research will address in the short term fine-tuning the architectures and algorithms and understanding their fundamental mathematical and psychophysical efficiencies, and in the long term issues such as multiple camera schemes and object based compression methods.
Q3. What is the basic approach to 3D-stereoscopic imagery?
APPROACHTheir basic approach to compression of 3D-stereoscopic imagery is based on the observation that disparity, the relative offset between corresponding points in an image pair, varies only slowly over most of the image field.
Q4. What is the effect of a stereoscopically viewed image?
When the set is as small (in bits) as 1 to 2% of the conventionally compressed image the stereoscopically viewed pair consisting of one original and one synthesized image produces convincing stereo imagery.
Q5. What are the main topics that the authors need to address in the context of 3D-stere?
Topics that the authors need to address in the context of compression of 3D-stereoscopic imagery include:• Optimizing implementation of the WorldLine approach.•
Q6. What is the way to exploit the high correlation between temporally adjacent frames?
The successful development of compression schemes for motion video that exploit the high correlation between temporally adjacent frames, e.g., MPEG, suggests that the authors might analogously exploit the high correlation between spatially or angularly adjacent still frames, i.e., left-right 3D-stereoscopic image pairs.
Q7. How do you compute predictors for left and right views?
Using three cameras: compute predictors for left and right views given the middle view, transmit the middle view and the predictors, synthesize 3D-stereoscopic views at the receiver.
Q8. What is the fundamental issue when 3D-stereoscopy is implemented on a?
The fundamental issue is that when 3D-stereoscopy is implemented on a single display each eye gets in some sense only half the display.
Q9. What is the effect of a synthesis of a left-right stereo image?
Their experiments demonstrate that a reasonable synthesis of one image of a left-right stereo image pair can be estimated from the other uncompressed or conventionally compressed image augmented by a small set of numbers that describe the local cross-correlations in terms of a disparity map.
Q10. What is the difference between the two views?
In fact, because the two views comprising a 3D-stereoscopic image pair are nearly identical, i.e., the information content of both together is only a little more than the information content of one alone, it is possible to find representations of image pairs and streams that take up little more storage space and transmission bandwidth than the space or bandwidth that is required by either alone.
Q11. How much bandwidth is needed to transmit 3D-stereoscopic images?
the bandwidth must apparently be doubled to transmit 3D-stereoscopic image streams at the same spatial resolution and temporal update frequency as either flat image stream.
Q12. What is the difference between the two components of a 3D-stereoscopic image?
One component may be either lossless or slightly lossy, as in conventional compression of flat imagery; the other component is by itself a very lossy (or "deep") method of compression.
Q13. What is the way to deal with occlusions?
The human visual perception system has an effective way to deal with occlusions: the authors have a detailed understanding of the image semantics, from which the authors effortlessly and unconsciously draw inferences that fill in the missing information.
Q14. What is the method for initial experiments?
This is the obvious candidate for initial experiments because it is easy to code and because the authors have a strong intuitive understanding of its parameters.
Q15. How does the price of a 3D-stereoscopic image be extracted?
The price may be extracted in either essentially the spatial domain, e.g., by assigning the odd lines to the left eye and the even lines to the right eye, or in essentially the temporal domain, e.g., by assigning alternate frames to the left and right eye.
Q16. How much is the net compression of disparity?
Each disparity is a vector with two components, horizontal and vertical, so the net compression has an upper bound of 1/32, about 3%.