Dynamic and Scalable Large Scale Image Reconstruction
SIFT-like local feature descriptors are ubiquitously employed in such computer vision applications as content-based retrieval, video analysis, copy detection, object recognition, photo-tourism and 3D reconstruction.
Feature descriptors can be designed to be invariant to certain classes of photometric and geometric transformations, in particular, affine and intensity scale transformations. However, real transformations that an image can undergo can only be approximately modeled in this way, and thus most descriptors are only approximately invariant in practice.
Secondly, descriptors are usually high-dimensional (e.g. SIFT is represented as a 128-dimensional vector). In large-scale retrieval and matching problems, this can pose challenges in storing and retrieving descriptor data.
We map the descriptor vectors into the Hamming space, in which the Hamming metric is used to compare the resulting representations.
This way, we reduce the size of the descriptors by representing them as short binary strings and learn descriptor invariance from examples.
We show extensive experimental validation, demonstrating the advantage of the proposed approach.
Venice internet image evaluation dataset: example patches corresponding to the same 3D point.
click to enlarge
click to enlarge
Confusion matrix for the 24 largest tracks for the Venice dataset.
SIFT L2 norm
LDAHash Hamming distance
click to enlarge
click to enlarge
Calibration results using the binary descriptors
Lausanne cathedral statue
Sequence contains 127 18-Megapixel images of a statue at different scales.
result of the calibration: camera positions and sparse 3D points
result of the calibration: 3D points (zoom)
click to enlarge
click to enlarge
Data set courtesy of M.Pauly.
download
images corrected for radial distortion [gz] for the original images, please contact christoph strecha
calibration: projection matrices [gz]
Lausanne cathedral side entry
Sequence contains 288 18-Megapixel images of the Lausanne cathedral side entry at different scales.
result of the calibration: camera positions and sparse 3D points
result of the calibration: 3D points (zoom)
click to enlarge
click to enlarge
Data set courtesy of M.Pauly.
download
images corrected for radial distortion [gz] for the original images, please contact christoph strecha
calibration: projection matrices [gz]
Venice
Sequence contains internet images of Venice at different scales and sizes.
result of the calibration: camera positions and sparse 3D points
result of the calibration: 3D points (zoom)
click to enlarge
click to enlarge
Data
We will be sharing the training data sets here in future. Stay tuned.
LDAHash: Improved matching with smaller descriptors
Christoph Strecha, Alex M. Bronstein, Michael M. Bronstein, Pascal Fua IEEE Transactions on Pattern Analysis and Machine Intelligence
Under Review Submitted August 2010
LDAHash: Improved matching with smaller descriptors
Christoph Strecha, Alex M. Bronstein, Michael M. Bronstein, Pascal Fua Techical Report pdf
|
@inproceedings{ldahashtr10,
author = "C. Strecha, A. M. Bronstein, M. M. Bronstein and Pascal Fua",
title = {{LDAHash: Improved matching with smaller descriptors}},
booktitle = "EPFL-REPORT-152487",
year = 2010,
}
Related References
Dynamic and Scalable Large Scale Image Reconstruction
C. Strecha, T. Pylvanainen, P. Fua CVPR 2010 website
| pdf
|
@inproceedings{ldahashtr10,
author = "C. Strecha, T. Pylvanainen, P. Fua",
title = {{Dynamic and Scalable Large Scale Image Reconstruction}},
booktitle = "Proceedings of 23rd {IEEE} Conference on Computer Vision and Pattern Recognition",
year = 2010,
}
On Benchmarking Camera Calibration and Multi-View Stereo for High Resolution Imagery
C. Strecha, W. von Hansen, L. Van Gool, P. Fua, U. Thoennessen CVPR 2010 website
| pdf
|
@inproceedings{ldahashtr10,
author = "C. Strecha, W. von Hansen, L. Van Gool, P. Fua, U. Thoennessen",
title = {{On Benchmarking Camera Calibration and Multi-View Stereo for High Resolution Imagery}},
booktitle = "Proceedings of 21rd {IEEE} Conference on Computer Vision and Pattern Recognition",
year = 2008,
}
Efficient Large Scale Multi-View Stereo for Ultra High Resolution Image Sets
Engin Tola, Christoph Strecha, Pascal Fua Machine Vision and Applications
Under Review Submitted September 2010 website
