Further exploration of the object-image metric with image registration in mind

Mendoza, Olga L.; Arnold, Gregory; Stiller, Peter F.

doi:10.1117/12.777181

Cited by 3 publications

(6 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The traditional image registration algorithms, such as the algorithm of Davis and Keck [71,72], try to register the two images by computing the homography matrix H between corresponding feature points. The limit of this algorithm is that they assume all the points in the physical world are coplanar or approximately coplanar, which is not true with high-rise scenes.…”

Section: Image Registration With Depth Informationmentioning

confidence: 99%

“…Once we obtain the 3D structure of the feature points, the motion, and calibration of the camera, we can start to register the rest of the pixels in the images with the estimated depth information. The (a) The 1st frame in the 'oldhousing' video sequence (b) The 88th frame in the 'oldhousing' video sequence Figure 75: Original frames used for image registration traditional image registration algorithms, such as the algorithm proposed by Davis and Keck [71,72], try to register the two images by computing the homography matrix H between corresponding feature points. The limit of this algorithm is that they assume all the points in the physical world are coplanar or approximately coplanar, which is not true with high-rise scenes.…”

Section: Image Registration With Depth Informationmentioning

confidence: 99%

“…Image registration is a fundamental task in image processing and computer vision which matches two or more images taken at different times and different viewpoints, by geometrically aligning reference and sensed images. There has been a broad range of techniques developed over the years The prevailing image registration methods, such as Davis and Keck's algorithm [71,72], assume all the feature points are coplanar and build a homography transform matrix to do registration.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Advanced Techniques for Scene Analysis

Wu¹

2010

View full text Add to dashboard Cite

The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.

show abstract

Section: Image Registration With Depth Informationmentioning

confidence: 99%

Section: Image Registration With Depth Informationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Advanced Techniques for Scene Analysis

Wu¹

2010

View full text Add to dashboard Cite

show abstract

“…Different techniques are compared for confidence, accuracy, timeliness, throughput, cost, and robustness. Image registration evaluation requires metrics [17], comparisons to image sets [18], and performance estimation and bounds [19]. Critical to IRM evaluation is a set of data for joint comparisons.…”

Section: Introductionmentioning

confidence: 99%

Video image registration evaluation for a layered sensing environment

Mendoza-Schrock

Patrick

Blasch

2009

Proceedings of the IEEE 2009 National Aerospace &Amp; Electronics Conference (NAECON)

View full text Add to dashboard Cite

In this paper, several methods to register and stabilize a motion imagery video sequence under the layered sensing concept are evaluated. Utilizing the layered sensing paradigm, an area is surveyed by a multitude of sensors at many different altitudes and operating across many modalities. Utilizing a combination of sensors provides better insight into a situation than could ever be achieved with a single sensor. A fundamental requirement in layered sensing is to first register, stabilize, and normalize the data from each of the individual sensors. This paper extends our previous work [1] to include experimental analysis. The paper contribution provides an evaluation of four registration algorithms now including the (1) Lucas-Kanade (LK) algorithm, (2) the Ohio State University (OSU) 1 correlation-based method, (3) robust data alignment (RDA), and (4) Scale Invariant Feature Transform (SIFT). Results demonstrate that registration accuracy and robustness were achieved with the LK and correlation-based methods over the others for image-to-image registration, restricted adaptive tuning, and stabilization over warped images; while the SIFT outperformed the others for partial image overlap.

show abstract

“…The prevailing image registration methods, such as the algorithm of Davis and Keck, 10,11 assume all the feature points are coplanar and build a homography transform matrix to do registration. The advantage is that they have low computational cost and can handle planar scenes conveniently; however, with the assumption that the scenes are approximately planar, they are inappropriate in the registration applications when the images have large depth variation due to the high-rise objects, known as the parallax problem.…”

Section: Introductionmentioning

confidence: 99%

Depth-based image registration

Han

Paulson

2010

SPIE Proceedings

View full text Add to dashboard Cite

Image registration is a fundamental task in computer vision because it can significantly contribute to high-level computer vision and benefit numerous practical applications. Though a lot of image registration techniques exist in literature, there is still a significant amount of research to be conducted because there are a lot of issues that need to be solved such as the parallax problem. The traditional image registration algorithms suffer from the parallax problem due to their underling assumption that the scene can be regarded approximately planar which is not satisfied in the case of large depth variation in the images with high-rise objects. With regard to the the parallax problem, a new strategy is proposed by leveraging the depth information via 3D reconstruction. One novel idea is to recover the depth in the image region with high-rise objects to build accurate transform function for image registration. Our method mitigates the parallax problem and can achieve robust registration results, which is validated by our experiments. Our algorithm is attractive to numerous practical applications.

show abstract

Further exploration of the object-image metric with image registration in mind

Cited by 3 publications

References 4 publications

Advanced Techniques for Scene Analysis

Advanced Techniques for Scene Analysis

Video image registration evaluation for a layered sensing environment

Depth-based image registration

Contact Info

Product

Resources

About