Fast normalized cross-correlation

2009 XXII Brazilian Symposium on Computer Graphics and Image Processing, 2009

This paper presents a novel methodology to perform matching between image points described by their respective features. Traditionally, such correspondences are determined by computing the similarity between descriptor vectors associated to each point which are obtained by invariant descriptors. Our methodology first obtains a coarse global registration among images, which constrains the correspondence space. Then, it analyzes the similarity among descriptors, thus reducing both the number and the severity of mismatches. The approach is sufficiently generic to be used with many feature descriptor methods. We present several experimental results that show significant increase in accuracy, number of successful matches, and execution time.

EURASIP Journal on Advances in Signal Processing, 2007

This paper presents a feature-list cross-correlation algorithm based on: a common feature extraction algorithm, a transformation of the results into a feature-list representation form, and a list-based cross-correlation algorithm. The feature-list cross-correlation algorithms are compared with known results of the common cross-correlation algorithms. Therefore, simple test images containing different objects under changing image conditions and with several image distortions are used. In addition, a medical application is used to verify the results. The results are analyzed by means of curve progression of coefficients and curve progression of peak signal-to-noise ratio (PSNR). As a result, the presented feature list cross-correlation algorithms are sensitive to all changes of image conditions. Therefore, it is possible to separate objects that are similar but not equal. Because of the high quantity of feature points and the strong PSNR, the loss of a few feature points does not have a significant influence on the detection results. These results are confirmed by a successfully applied medical application. The calculation time of the feature list cross-correlation algorithms only depends on the length of the feature-lists. The amount of feature points is much less than the number of pixels in the image. Therefore, the feature-list cross-correlation algorithms are faster than common cross-correlation algorithms. Better image conditions tend to reduce the size of the feature-list. Hence, the processing time decreases considerably.

Proceedings of the 23rd Spring Conference on Computer Graphics - SCCG '07, 2007

Feature matching is based on finding reliable corresponding points in the images. This requires to solve a twofold problem: detecting repeatable feature points and describing them as distinctive as possible. SIFT (Scale Invariant Feature Transform) has been proven to be the most reliable solution to this problem. It combines a scale invariant detector and a very robust descriptor based on gray image gradients. Even if in general the detected SIFT feature points have a repeatability score greater than 40 %, an important proportion of them are not identified as good corresponding points by the SIFT matching procedure. In this paper we introduce a new and effective method that increases the number of valid corresponding points. To improve the distinctness of the original SIFT descriptor the color information is considered. In our method we compute the cross correlation and the histogram intersection between neighbour keypoints regions that has been adapted to scale and rotation. Finally, the experimental results prove that our method outperforms the original matching method.

Pattern Recognition, 2000

The main aim of this work is to propose a new technique to solve the well-known feature correspondence problem for motion estimation. The problem is formulated as an optimization process whose energy function includes constraints based on projective invariance of cross-ratio of "ve coplanar points. Starting from some approximated correspondences, estimated by radiometric similarity, for features with high directional variance, optimal matches are obtained through an optimization technique. The new contribution of this work consists of a matching process, re"ning the raw measurements, based on an energy function minimization technique converging to an optimal solution for most of the features by taking advantage of some good initial guess, and in the use of cross ratio as geometrical invariant constraint to detect and correct the mismatches due to wrong radiometric similarity measures. Though the method is based on geometrical invariance of coplanar points, it is not required that all features have to be coplanar or to preprocess the images to detect the planar regions. Experimental results are presented for real and synthetic images, and the performance of the novel approach is evaluated on di!erent image sequences and compared to well-known techniques.

Communications in Computer and Information Science, 2011

Cross-correlation is an important image processing algorithm for template matching widely used on computer vision based systems. This work follows a profile-based hardware/software co-design method to develop an architecture for normalized cross-correlation coefficient calculus using Nios II soft-processor. Results present comparisons between general purpose processor implementation and different customized softprocessor implementation considering execution time and the influence of image and sub-image size. Nios II soft-processors configured with floating-point hardware acceleration achieved a 8.31 speedup.

B. M a t h a n k u m a r , N. S u r e s h k u m a r & T. C h i t r a k u m a r Abstract:-Image search in image compression includes the mapping of visual features into compact binary codes. It needs more memory space for storage. So in order to reduce the memory cost for storage, there is a need for encoding high-dimensional data as compact binary codes. The similarity is also an important consideration in large scale image search. The computational efficiency can be computed based on the computation of similarity. It is measured by using hamming distance. Signals are used to detect the feature. It is achieved by taking correlation between two signals (cross correlation).For high frequency Domain expression the normalized form of correlation (correlation coefficient) preferred in template matching does not have a correspondingly simple and efficient frequency domain expression. To avoid this normalized cross-correlation has been computed in the spatial domain. As the computational cost of spatial domain convolution is high, more inexact but fast spatial domain matching methods have also been developed. To obtain normalized cross correlation from transform domain convolution the Fast Normalized Cross Correlation Algorithm is introduced. This new algorithm provides an order of magnitude speedup over spatial domain computation of normalized cross correlation.

2002

This paper proposes an empirical evaluation of different matching strategies that have been proposed in the literature to solve the problem of feature point correspondence between images. They will be evaluated in terms of their ability to reduce the number of false matches in given match sets, while preserving the good matches. The validation process determines the number of good matches and the proportion of good matches in a given match set, and this for the different parameter values of a matching constraint.

The development of optoelectronics and increasing the processing speed of processors encourage the use of different measurement methods with the use of data obtained from the digital cameras. The article presents the test results of the time consumption of the normalised 2D-correlation function. The correlation is one of most commonly used functions in image processing. It searches for a correlation between the two images and requires quite a lot of calculation operations. The speed of this function was examined in a variety of tools, such as MatLab, MatLab with tools Parallel Computing, library OpenCV and OpenCV wer.2.2 with the CUDA platform extension. During tests, the speeds of image processing were determined, especially for images acquired with high-resolution cameras. Conclusions presented concern with application possibilities of correlation methods, including multithreaded processing.

Matching images is a fundamental problem in image processing. The most common technique used to compare binary images is to calculate the correlation between two images or simply to subtract them. Both of these methods –as well as other matching methods– require some type of similarity operation to be applied to the whole image, and hence they are image size dependent. This implies that as image size increases, more processing time is required. However, with image sizes already exceeding 20 mega-pixels and standard image sizes doubling approximately every five years, the need to find a size invariant image matching method is becoming crucial. In this paper, we present a quick way to compare and match binary images based on the Probabilistic Matching Model (PMM). We present two simple image size invariant methods based on PMM: one for fast detection of dissimilar binary images and another for matching binary images. For detecting dissimilar binary images we introduce the Dissimilar Detection via Mapping method (DDM). We compare DDM to other popular matching methods used in the image processing arena and show that DDM is magnitudes faster than any other method. For binary image matching, we use DDM as a preprocessor for other popular methods to speed up their matching speed. In particular, we use DDM with cross correlation to speed it up. Test results are presented for real images varying in size from 16 kilo-pixel images to 10 mega-pixel images to show the method's size invariance.

IEEE Transactions on Medical Imaging, 2011