A Comparison of Two Typical Local Feature Matching Algorithm: SIFT and MSER

2014 ◽  
Vol 687-691 ◽  
pp. 4119-4122
Author(s):  
Xiao Cun Jiang ◽  
Xiao Liu ◽  
Tao Tang ◽  
Xiao Hu Fan ◽  
Xiao Cui
Keyword(s):  
Feature Matching ◽  
Local Feature ◽  
Scale Invariant Feature Transform ◽  
Scale Invariant ◽  
Matching Algorithm ◽  
Invariant Feature ◽  
Distance Difference ◽  
Feature Transform ◽  
Scale Invariant Feature ◽  
Shooting Distance

Scale invariant feature transform matching algorithm and Maximally Stable Extremal Regions matching algorithm have been widely used because of their good performance. The two local feature matching algorithms were compared through numbers of experiments in this paper. The experiment results showed that SIFT is good at dealing with the image distortion from shooting distance difference and small shooting viewpoint deviation; MSER is good at handling the complicated affine distortion from big shooting viewpoint deviation. From the aspect of scene types, the performance of SIFT is good both to structure images and texture images. MSER is suitable for the matching of structure images, but not so successful to that of texture images.

scholarly journals Bidirectional scale-invariant feature transform feature matching algorithms based on priority k-d tree search

2016 ◽  
Vol 14 (1) ◽  
pp. 172988141668270
Author(s):  
XiangShao Liu ◽  
Shangbo Zhou ◽  
Hua Li ◽  
Kun Li
Keyword(s):  
Image Registration ◽  
Feature Matching ◽  
Scale Invariant Feature Transform ◽  
Tree Search ◽  
Scale Invariant ◽  
Matching Algorithm ◽  
Invariant Feature ◽  
Feature Transform ◽  
Key Steps ◽  
Scale Invariant Feature

In this article, a bidirectional feature matching algorithm and two extended algorithms based on the priority k-d tree search are presented for the image registration using scale-invariant feature transform features. When matching precision of image registration is below 50%, the discarding wrong match performance of many robust fitting methods like Random Sample Consensus (RANSAC) is poor. Therefore, improving matching precision is a significant work. Generally, a feature matching algorithm is used once in the image registration system. We propose a bidirectional algorithm that utilizes the priority k-d tree search twice to improve matching precision. There are two key steps in the bidirectional algorithm. According to the case of adopting the ratio restriction of distances in the two key steps, we further propose two extended bidirectional algorithms. Experiments demonstrate that there are some special properties of these three bidirectional algorithms, and the two extended algorithms can achieve higher precisions than previous feature matching algorithms.

An Improved SIFT Matching Algorithm

2012 ◽  
Vol 239-240 ◽  
pp. 1232-1237 ◽  
Author(s):  
Can Ding ◽  
Chang Wen Qu ◽  
Feng Su
Keyword(s):  
High Dimension ◽  
Gradient Method ◽  
Feature Matching ◽  
Feature Descriptor ◽  
Scale Invariant ◽  
Matching Algorithm ◽  
Invariant Feature ◽  
Feature Transform ◽  
New Feature ◽  
Scale Invariant Feature

The high dimension and complexity of feature descriptor of Scale Invariant Feature Transform (SIFT), not only occupy the memory spaces, but also influence the speed of feature matching. We adopt the statistic feature point’s neighbor gradient method, the local statistic area is constructed by 8 concentric square ring feature of points-centered, compute gradient of these pixels, and statistic gradient accumulated value of 8 directions, and then descending sort them, at last normalize them. The new feature descriptor descend dimension of feature from 128 to 64, the proposed method can improve matching speed and keep matching precision at the same time.

scholarly journals A non-target structural displacement measurement method using advanced feature matching strategy

2019 ◽  
Vol 22 (16) ◽  
pp. 3461-3472 ◽  
Author(s):  
Chuan-Zhi Dong ◽  
F Necati Catbas
Keyword(s):  
Measurement Method ◽  
Feature Matching ◽  
Displacement Measurement ◽  
Scale Invariant Feature Transform ◽  
Scale Invariant ◽  
Invariant Feature ◽  
Feature Transform ◽  
Matching Strategy ◽  
Scale Invariant Feature ◽  
Visual Geometry

Most of the existing vision-based displacement measurement methods require manual speckles or targets to improve the measurement performance in non-stationary imagery environments. To minimize the use of manual speckles and targets, feature points regarded as virtual markers can be utilized for non-target measurement. In this study, an advanced feature matching strategy is presented, which replaces the handcrafted descriptors with learned descriptors called Visual Geometry Group, of the University of Oxford descriptors to achieve better performance. The feasibility and performance of the proposed method is verified by comparative studies with a laboratory experiment on a two-span bridge model and then with a field application on a railway bridge. The proposed approach of integrated use of Scale Invariant Feature Transform and Visual Geometry Group improved the measurement accuracy by about 24% when compared with the commonly used existing feature matching-based displacement measurement method using Scale Invariant Feature Transform feature and descriptor.

scholarly journals Classification of wood knots using artificial neural networks with texture and local feature-based image descriptors

Holzforschung ◽  
2021 ◽  
Vol 0 (0) ◽  
Author(s):  
Sung-Wook Hwang ◽  
Taekyeong Lee ◽  
Hyunbin Kim ◽  
Hyunwoo Chung ◽  
Jong Gyu Choi ◽  
...  
Keyword(s):  
Local Feature ◽  
Scale Invariant Feature Transform ◽  
Scale Invariant ◽  
Invariant Feature ◽  
Feature Based ◽  
Feature Transform ◽  
Artificial Neural ◽  
Occurrence Matrix ◽  
Scale Invariant Feature

Abstract This paper describes feature-based techniques for wood knot classification. For automated classification of macroscopic wood knot images, models were established using artificial neural networks with texture and local feature descriptors, and the performances of feature extraction algorithms were compared. Classification models trained with texture descriptors, gray-level co-occurrence matrix and local binary pattern, achieved better performance than those trained with local feature descriptors, scale-invariant feature transform and dense scale-invariant feature transform. Hence, it was confirmed that wood knot classification was more appropriate for texture classification rather than an approach based on morphological classification. The gray-level co-occurrence matrix produced the highest F1 score despite representing images with relatively low-dimensional feature vectors. The scale-invariant feature transform algorithm could not detect a sufficient number of features from the knot images; hence, the histogram of oriented gradients and dense scale-invariant feature transform algorithms that describe the entire image were better for wood knot classification. The artificial neural network model provided better classification performance than the support vector machine and k-nearest neighbor models, which suggests the suitability of the nonlinear classification model for wood knot classification.

Sign in / Sign up

Export Citation Format

Share Document