SURF Method Research Articles

An Analysis of the SURF Method

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.

Comparing Proposed Signature with SURF in Object Detection Process

ABSTRACTMost object detection and classification algorithms only locate regions in the image, whether they are within a template-sliding mask or in interested region blobs. However, such regions may be ambiguous, especially when the object of interest is very small or unclear. This paper proposes an algorithm for automatic object detection and matching based on its own signature using morphological segmentation tools. Moreover, the algorithm tries to match objects, not in object blobs or regions of interest, but among the constructed proposed object signatures. During the matching process, the SURF method makes a comparison of process between its matching performance and the proposed matching process on all experimental objects. The process has tested a wide variety of 120 dissimilar objects; it has achieved 100% of constructed signatures, and it has achieved 96% of correct object matching; whereas SURF has achieved only 85% for all tested objects.

Proposed Method for Detecting Objects

of object detection and classification algorithms are only locating regions in the image, whether it is within a template- sliding mask or interested region blobs. However, such regions may be ambiguous, especially when the object of interest is very small, unclear, or anything else. This paper presents proposed algorithm for automatic object detection and matching based on its own proposed signature using morphological segmentation tools. Moreover, the algorithm tries to match the objects; neither among object's blobs nor among regions of interest; but among the constructed proposed objects' signatures. During the matching process, SURF method has presented to make a comparison of the experimental results. The performance has been tested 120 from a wide variety of unlike objects; it has been achieved 100% in the case of constructing object signatures, also it has been achieved 96% of right matching whereas SURF has achieved 85% for all test objects.

Water Hazard Detection for Intelligent Vehicle Based on Vision Information

Water hazard is one of the most dangerous obstacles for unmanned vehicle to travel in off-road environment. The vision information for water hazard body includes relatively high brightness, low saturation and smooth texture. In this paper, static image and sequence image are captured by CCD. Color and texture information extracted from image pixels would be utilized to form a feature matrix to train the SVM classifier. While, with the assistance of RBF kernel function, low-dimensional sample space is projected to high-dimensional space. Based on large amount of experiments, RBF kernel function parameters are optimized by using grid method. The optimized RBF kernel function parameters are proved satisfactory when detecting in the static water image. Notice that there is certain relationship between target position and its scale for the adjacent frame. SURF feature detection and matching method can be used to match feature point between adjacent frames in image sequences. The searching window size and position in new frame could be updated in time, which allows to detect water hazard in a relatively small region. Water obstacle tracking experiment with patents proved the satisfactory performance of the SURF method in this paper.

FEATURE RECOGNITION BERBASIS CORNER DETECTION DENGAN METODE FAST, SURF DAN FLANN TREE UNTUK IDENTIFIKASI LOGO PADA AUGMENTED REALITY MOBILE SYSTEM

Logo is a graphical symbol that is the identity of an organization, institution, or company. Logo is generally used to introduce to the public the existence of an organization, institution, or company. Through the existence of an agency logo can be seen by the public. Feature recognition is one of the processes that exist within an augmented reality system. One of uses augmented reality is able to recognize the identity of the logo through a camera.The first step to make a process of feature recognition is through the corner detection. Incorporation of several method such as FAST, SURF, and FLANN TREE for the feature detection process based corner detection feature matching up process, will have the better ability to detect the presence of a logo. Additionally when running the feature extraction process there are several issues that arise as scale invariant feature and rotation invariant feature. In this study the research object in the form of logo to the priority to make the process of feature recognition. FAST, SURF, and FLANN TREE method will detection logo with scale invariant feature and rotation invariant feature conditions. Obtained from this study will demonstration the accuracy from FAST, SURF, and FLANN TREE methods to solve the scale invariant and rotation invariant feature problems.

Matching of Feature Points Based on TSSC Method from MR Images of Nonrigid Deformed Tissues

Due to the nonlinear and nonuniform local deformation of nonrigid tissues, it is difficult to match a number of feature points distributing somewhat uniform in the tissues from MR images for deformation measurement. This paper proposes TSSC (TPS-SURF-SAC-Clustering) based method of feature point matching and elimination of mismatching. First, Fast-Hessian and Harris operator are utilized to extract the feature points in the initial MR image, and the matching region is identified by TPS transformation model for every query point in the deformed image. Then the SURF descriptors and the proposed Spatial Association Correspondence (SAC) method are combined to match the feature points. Finally, by clustering the coordinate differences between the matching points obtained by TPS-SURF-SAC and the matching points calculated by TPS model, most incorrectly matched points are eliminated. After every iterative processing of matching and mismatching elimination, the updated TPS model becomes more accurate and more correctly so that the matched points can be identified than those of last iteration. The experimental results show that the proposed SAC was efficient and that TSSC based method outperformed the single SURF or SIFT method.

CAMSHIFT IMPROVEMENT WITH MEAN-SHIFT SEGMENTATION, REGION GROWING, AND SURF METHOD

CAMSHIFT algorithm has been widely used in object tracking. CAMSHIFT utilizescolor features as the model object. Thus, original CAMSHIFT may fail when the object color issimilar with the background color. In this study, we propose CAMSHIFT tracker combined withmean-shift segmentation, region growing, and SURF in order to improve the tracking accuracy.The mean-shift segmentation and region growing are applied in object localization phase to extractthe important parts of the object. Hue-distance, saturation, and value are used to calculate theBhattacharyya distance to judge whether the tracked object is lost. Once the object is judged lost,SURF is used to find the lost object, and CAMSHIFT can retrack the object. The Object trackingsystem is built with OpenCV. Some measurements of accuracy have done using frame-basedmetrics. We use datasets BoBoT (Bonn Benchmark on Tracking) to measure accuracy of thesystem. The results demonstrate that CAMSHIFT combined with mean-shift segmentation, regiongrowing, and SURF method has higher accuracy than the previous methods.

A Case of Peripheral Artery Disease with Ipsilateral External Iliac and Superficial Femoral Artery Lesions Which Could Treated by Bidirectional Approach from Ipsilateral Femoral Artery Using SURF (Sidewinder<sup>®</sup> Catheter Used Reverse Femoral Sheath) Method

A Case of Peripheral Artery Disease with Ipsilateral External Iliac and Superficial Femoral Artery Lesions Which Could Treated by Bidirectional Approach from Ipsilateral Femoral Artery Using SURF (Sidewinder<sup>®</sup> Catheter Used Reverse Femoral Sheath) Method

고속 교통표시판 인식 알고리즘

자동차가 IT 기술과 융합되면서 편의성과 안전성 그리고 성능이 좋아지고 있다. 이와 관련하여 최근 자동차의 주행시 안전 및 주변 환경과 관련된 정보를 제공하기 위한 많은 알고리즘이 연구되고 있으며 교통표지판 인식 또한 그 중 하나이다. 교통표지판 인식은 안전 운전에 필요한 중요한 정보를 제공해 준다. 본 논문에서는 연산 시간 감소에 중점을 두어 교통표지판을 탐지하고 판별하는 인식 알고리즘을 제안한다. 제안한 알고리즘에서는 색상 임계값을 이용하여 교통표지판 후보를 분할하고 다각형 근사법을 이용하여 적절한 다각형을 찾는다. 이렇게 찾은 패턴에 대해 SURF와 ORB 알고리즘을 이용하여 데이터베이스와 비교하여 교통표지판을 식별한다. Information technology improves convenience, safety, and performance of automobiles. Recently, a lot of algorithms are studied to provide safety and environment information for driving, and traffic sign recognition is one of them. It can provide important information for safety driving. In this paper, we propose a method for traffic sign detection and identification concentrating on reducing the computation time. First, potential traffic signs are segmented by color threshold, and a polygon approximation algorithm is used to detect appropriate polygons. The potential signs are compared with the template signs in the database using SURF and ORB feature matching method.

모바일 시각 검색 시스템의 성능 향상을 위하여 개선된 Speeded Up Robust Features(SURF) 알고리듬

본 논문에서는 지역 특징 추출 방법 중 하나인 SURF를 개선시켜 모바일 환경에 적합한 특징 추출 및 정합 방법에 대하여 기술한다. 모바일 환경에서 최적의 성능을 내기 위해 크게 세 가지 방법을 제안한다. 첫 번째는 SURF 기술자의 차원을 축소시키는 방법이다. 기존의 64차원과 여러 차원의 기술자들의 성능을 비교 분석한다. 두 번째로 고속 헤시안 검출기에서 헤시안 행렬의 대각합을 통해 계산되는 라플라시안 부호를 사용하여 정합 성능을 높인다. 즉, 부호가 같으면 같은 특성의 특징점으로 여겨 정합하고 그렇지 않으면 정합하지 않는다. 세 번째는 특징점간의 거리 측정 시 사용하는 거리 비율을 실험하고, 실험을 통해 얻어진 거리 비율을 사용하여 정합의 성능을 높였다. 마지막으로, 기존의 시스템과 제안한 시스템의 성능을 비교하여 제안한 시스템의 성능 향상을 검증하고, 제안한 시스템이 비교적 정합률의 성능 저하 없이 응답속도를 크게 개선함을 보였다. In the paper, we propose enhanced feature extraction and matching methods for a mobile environment based on modified SURF. We propose three methods to reduce the computational complexity in a mobile environment. The first is to reduce the dimensions of the SURF descriptor. We compare the performance of existing 64-dimensional SURF with several other dimensional SURFs. The second is to improve the performance using the sign of the trace of the Hessian matrix. In other words, feature points are considered as matched if they have the same sign for the trace of the Hessian matrix, otherwise considered not matched. The last one is to find the best distance-ratio which is used to determine the matching points. We find the best distance-ratio through experiments, and it gives the relatively high accuracy. Finally, existing system which is based on normal SURF method is compared with our proposed system which is based on these three proposed methods. We present that our proposed system shows reduced response time while preserving reasonably good matching accuracy.

Human Pose Tracking For Video Using SURF features

In this paper, a novel method based on SURF features method for tracking human motion in monocular videos is proposed. With a initial human skeleton joint point template, we use the probability density propagation of the particle filers through the model. This algorithm can automatically achieve right human motion figure from tracking failures, such as occlusion and auto-occlusion problem. Experimental results from 20 classes monocular videos show that the new Based on SURF method is robust and the tracking results are good.

Construction of systems of local invariant image indicators based on the Fourier-Mellin transform

Two approaches to constructing systems of local invariant image indicators are compared: the SURF method and a method based on the incomplete Fourier-Mellin transform. The results for images made indoors show that the two methods are of comparable quality.