Modeling of three-dimensional (3D) objects from image sequences is a challenging problem and has been a research topic for many years. Important theoretical and algorithmic results were achieved that allow to extract even complex 3D models of scenes from sequences of images. One recent effort has been to reduce the amount of calibration and to avoid restrictions on the camera motion. In this contribution an approach is described which achieves this goal by combining state-of-the-art algorithms for uncalibrated projective reconstruction, self-calibration and dense correspondence matching.
In applications where nearly reversible color image compression is required, the choice of an appropriate color space is a major factor determining the attainable compression ratio.The optimal choice would satisfy the following conditions . Decorrelate the data as much as possible.. Minimize the total number of bits in the data path needed for a certain quality level (assuming a truly reversible component compression scheme). a Need no calculations with critical accuracy.We compared several known linear color transformations (under the assumption that the original data was represented in RGB) with respect to these conditions. Decorrelation and entropy reduction were calculated. The number of bits required to guarantee a certain quality level measured in SLab units was determined. The effects of transformation coefficient quantization were checked.As a result, a simple transformation needing no multipliers at all is proposed.
To reach even still higher compression ratios in video sequence coding the demands posed on the motion estimation module become higher and higher. In this paper we will present a motion estimation scheme which gives a motion field defined on every block (typical 8 by 8 or 16 by 16) and resembles very well the real motion in the scene. The first idea behind motion compensated prediction was just to minimize the local energy of a given block so that it could efficiently be coded. In more complex coding schemes other constraints were added to obtain better visual quality. The most often used constraint is the one to have a smooth motion field which is felt to have a higher correspondence to the real motion compared to the rather "noisy" motion fields obtained from the classical MAD schemes. Many strategies exist to get to these "smooth" motion fields. Most of them are based on some statistical modeling of the resulting motion vector field which asks for iterative solutions. The proposed algorithm is a "one-pass" algorithm and is not explicitly based on any statistical model. A classification procedure on a predefined number of motion vector candidates defines the final motion field.Applications for such a type of motion estimation algorithms can be found in very low bitrate coding where the amount of bits to code the motion field is important compared to the total bitrate , in codecs which want to exploit some features of the human visual system (classification for optimal bit allocation, use of motion masking. . .) or in object based coding where the motion estimation algorithm interacts with the segmentation procedure.In a first section an overview will be given of the concept of using global information to calculate the motion vector field. The second part gives a more thorough description of the new algorithm and in the last part some simulation results will be presented on real sequences. Further research on this algorithm will be done to use it in a segmentation based codec for complex scenes.
In video coding it is clearly worthwhile to have a more realistic motion field than what can be obtained by the classical mean absolute difference (MAD) full search method. Applications can be found in very low bitrate coding where the amount of bits to code the motion field is important compared to the total bitrate, in codecs which want to exploit some features of the human visual system (classification for optimal bit allocation, use of motion masking. ..) or in object based coding where the motion estimation algorithm interacts with the segmentation procedure. Our research started from the well known MAD full search procedure and wanted to obtain reasonable results without adding too much complexity. The improvement is performed inside the algorithm without any need for post processing. After a more thorough description of these improvements, some results will be compared and applications indicated. 1. DIFFERENCES WITH FULL SEARCH MAD There are several reasons why a full search MAD often results in vectors that are not indicating real motion. SPIEVo!. 1977/251 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with đŸ’™ for researchers
Part of the Research Solutions Family.