An Analysis of the SURF Method

Abstract

The SURF method (Speeded Up Robust Features) is a fast and robust algorithm for local, similarity invariant representation and comparison of images. Similarly to many other local descriptor-based approaches, interest points of a given image are defined as salient features from a scale-invariant representation. Such a multiple-scale analysis is provided by the convolution of the initial image with discrete kernels at several scales (box filters). The second step consists in building orientation invariant descriptors, by using local gradient statistics (intensity and orientation). The main interest of the SURF approach lies in its fast computation of operators using box filters, thus enabling real-time applications such as tracking and object recognition. The SURF framework described in this paper is based on the PhD thesis of H. Bay [ETH Zurich, 2009], and more specifically on the paper co-written by H. Bay, A. Ess, T. Tuytelaars and L. Van Gool [Computer Vision and Image Understanding, 110 (2008), pp. 346-359]. An implementation is proposed and used to illustrate the approach for image matching. A short comparison with a state-of-the-art approach is also presented, the SIFT algorithm of D. Lowe [International Journal of Computer Vision, 60 (2004), pp. 91-110], with which SURF shares a lot in common. Source CodeThe source code and the online demo are accessible at the IPOL web page of this article 1 . The proposed implementation of the SURF algorithm is written in C++ ISO/ANSI. It performs features extraction from digital images and provides local correspondences for a pair of images. An epipolar geometric consistency checking may additionally be used to discard mismatches when considering two pictures from the same scene. This optional post-processing uses the ORSA algorithm by B.

A Survey on Disease Detection in Plant using Image Processing Techniques

“…In image acquisition acquired the image then removes noise. The next step segmentation is to classify and modify rendering of an image then extracting the feature using speed up robust feature (SURF) methods [17] is useful for extracting the features and last step classification is done ANN. Bagde et al [18] explained artificial neural network-based plant leaf disease identification".…”

Evaluation of Local Matching Methods in Image Analysis for Mineral Grain Tracking in Microscope Images of Rock Sections

“…An image search under different enlargements, with scaling performed by computer methods, rendered satisfactory results only for the SURF method; however, it was sensitive to significant data reduction. SURF ensures that the points of interest are scale-invariant by transforming the image using the multi-resolution pyramid technique, to copy the original image with a Gaussian pyramid or Laplacian pyramid shape, in order to obtain an image with a special blurring effect on the original image, called scale-space [29,30]. Image acquisition with different microscope magnifications is characterized by different, resolutions of a pixel, therefore resulting in different representations of the values of the pixels neighbouring the searched image.…”

Image stack alignment in full-field X-ray absorption spectroscopy using SIFT_PyOCL

“…The SIFT algorithm (Lowe, 1999(Lowe, , 2004, widely used for panoramic image stitching, has been adapted to FFXAS from the IPOL implementation (Yu & Morel, 2011). Another registration algorithm, SURF [for speeded up robust feature (Bay et al, 2008;Oyallon & Rabin, 2013)], has been evaluated; it produces fewer keypoints with a keypoint descriptor twice as small as that generated by SIFT (64 bytes instead of 128). While being faster than SIFT, this algorithm was not retained due to some coarse approximation in the blurring procedure (box-filtering) and its inadequate descriptor size, making matching less reliable.…”