Scene Matching for Infrared and Visible Images with Compressive Sensing SIFT Feature Representation in Bandelet Domain

2020 ◽  
Vol 34 (11) ◽  
pp. 2054029
Author(s):  
Gang Liu ◽  
Sen Liu ◽  
Liuke Liang ◽  
Zhonghua Liu ◽  
Jianwei Ma
Keyword(s):  
Compressive Sensing ◽  
Distance Measure ◽  
Infrared Image ◽  
Low Frequency ◽  
Similarity Criterion ◽  
Feature Representation ◽  
Geometric Distortion ◽  
Matching Problem ◽  
Sift Feature ◽  
Scene Matching

Aimed at scene matching problem for taking infrared image as the actual data and the visible image as the referenced data, a multi-resolution matching algorithm which fuses compressive sensing Scale Invariant Feature Transform (SIFT) feature is presented based on Bandelet transform. Two kinds of images are separately transformed into Bandelet domain to compress the feature search space of scene matching based on the best sparse representation of natural images by Bandelet transform. On the basis of adaptive Bayes threshold denoising for infrared image, the concept of sparse feature representation of compressive sensing theory is introduced into SIFT algorithm. For low-frequency image in Bandelet domain, high-dimensional SIFT key point feature description vector is projected on compressive sensing random measurement matrix to achieve dimensionality reduction. Then, the improved Genetic Algorithm (GA) to overcome premature phenomena is used as the search strategy, and the L1 distance measure of SIFT feature vectors of compressive sensing for two kinds of images is applied to the search similarity criterion to match low-frequency image of high scale in Bandelet domain. The matching result is used as the guidance of the matching process for low-frequency image of low scale, and the matching result of full-resolution image is obtained iteratively. Experimental results show that the proposed method has not only high matching accuracy and fast matching speed, but also better robustness in comparison with some classic matching algorithms, which can resist the geometric distortion of rotation for actual image.

scholarly journals Magnetic Normal Mode Calculations in Big Systems: A Highly Scalable Dynamical Matrix Approach Applied to a Fibonacci-Distorted Artificial Spin Ice

2021 ◽  
Vol 7 (3) ◽  
pp. 34
Author(s):  
Loris Giovannini ◽  
Barry W. Farmer ◽  
Justin S. Woods ◽  
Ali Frotanpour ◽  
Lance E. De Long ◽  
...  
Keyword(s):  
Normal Modes ◽  
Low Frequency ◽  
Fibonacci Sequence ◽  
Exchange Interactions ◽  
Geometric Distortion ◽  
Dynamical Matrix ◽  
Spin Ice ◽  
Magnetization Dynamics ◽  
Distortion Parameter ◽  
Artificial Spin Ice

We present a new formulation of the dynamical matrix method for computing the magnetic normal modes of a large system, resulting in a highly scalable approach. The motion equation, which takes into account external field, dipolar and ferromagnetic exchange interactions, is rewritten in the form of a generalized eigenvalue problem without any additional approximation. For its numerical implementation several solvers have been explored, along with preconditioning methods. This reformulation was conceived to extend the study of magnetization dynamics to a broader class of finer-mesh systems, such as three-dimensional, irregular or defective structures, which in recent times raised the interest among researchers. To test its effectiveness, we applied the method to investigate the magnetization dynamics of a hexagonal artificial spin-ice as a function of a geometric distortion parameter following the Fibonacci sequence. We found several important features characterizing the low frequency spin modes as the geometric distortion is gradually increased.

Research on Adaptive Face Gender Recognition Based on Compressive Sensing

2014 ◽  
Vol 989-994 ◽  
pp. 4187-4190 ◽  
Author(s):  
Lin Zhang
Keyword(s):  
Feature Extraction ◽  
Compressive Sensing ◽  
Recognition Accuracy ◽  
Recognition Rate ◽  
Low Frequency ◽  
Face Image ◽  
Gender Recognition ◽  
Recognition Method ◽  
Frequency Information ◽  
Learning Machine

An adaptive gender recognition method is proposed in this paper. At first, do multiwavlet transform to face image and get its low frequency information, then do feature extraction to the low frequency information using compressive sensing (CS), use extreme learning machine (ELM) to achieve gender recognition finally. In the process of feature extraction, we use genetic algorithm (GA) to get the number of measurements of CS in order to gain the highest recognition rate, so the method can adaptive access optimal performance. Experimental results show that compared with PDA and LDA, the new method improved the recognition accuracy substantially.

scholarly journals Flying Small Target Detection for Anti-UAV Based on a Gaussian Mixture Model in a Compressive Sensing Domain

Sensors ◽  
2019 ◽  
Vol 19 (9) ◽  
pp. 2168 ◽  
Author(s):  
Chuanyun Wang ◽  
Tian Wang ◽  
Ershen Wang ◽  
Enyan Sun ◽  
Zhen Luo
Keyword(s):  
Compressive Sensing ◽  
Target Detection ◽  
Sparse Matrix ◽  
Infrared Image ◽  
Matrix Decomposition ◽  
Gaussian Mixture ◽  
Visual Surveillance ◽  
Low Rank ◽  
Small Target ◽  
Small Target Detection

Addressing the problems of visual surveillance for anti-UAV, a new flying small target detection method is proposed based on Gaussian mixture background modeling in a compressive sensing domain and low-rank and sparse matrix decomposition of local image. First of all, images captured by stationary visual sensors are broken into patches and the candidate patches which perhaps contain targets are identified by using a Gaussian mixture background model in a compressive sensing domain. Subsequently, the candidate patches within a finite time period are separated into background images and target images by low-rank and sparse matrix decomposition. Finally, flying small target detection is achieved over separated target images by threshold segmentation. The experiment results using visible and infrared image sequences of flying UAV demonstrate that the proposed methods have effective detection performance and outperform the baseline methods in precision and recall evaluation.

A New Infrared Image Magnification and Enhancement Method Based on Wavelet Transform

2012 ◽  
Vol 198-199 ◽  
pp. 238-243 ◽  
Author(s):  
Wen Sheng Guo ◽  
Feng Chen ◽  
Zhao You Sun ◽  
Xi Jun Wang
Keyword(s):  
Wavelet Transform ◽  
High Frequency ◽  
Infrared Image ◽  
Low Frequency ◽  
High Energy ◽  
Low Energy ◽  
Image Magnification ◽  
Edge Profile ◽  
Enhancement Method ◽  
Image Edge

The traditional image magnify method usually have some defects on details. This paper gives a new infrared image magnification and enhancement method which is based on wavelet reconstruction and gradation segment. In this method, first of all, make wavelet transform on the image, get the high-frequency coefficient. Apply the Newton differential algorithm enhance the high-frequency coefficient as the high-frequency part of the magnified image, treat the original image as the low-frequency part , make the wavelet reconstruction ,then get the magnified image. To enhance the magnified image, according to the double gray threshold, segment the image into high gray segment corresponding to target, low gray segment corresponding to background, and middle gray segment corresponding to transition sector. Then, make linear extension to them respectively; the result is the magnified image. Experiments indicate, this method is effective on distinguishing high-energy target from low-energy target (the low-energy target is the primary one) and displaying the details of image(edge profile of the bomb).

Method of Footprint Image Stitching Based on Multiscale SIFT Feature Matching

2013 ◽  
Vol 303-306 ◽  
pp. 1056-1059
Author(s):  
Sen Wang ◽  
Yin Hui Zhang ◽  
Zhong Hai Shi ◽  
Zi Fen He
Keyword(s):  
Wavelet Transform ◽  
Feature Matching ◽  
Wavelet Coefficient ◽  
Low Frequency ◽  
Original Method ◽  
Inverse Transformation ◽  
Image Stitching ◽  
Image Splicing ◽  
Sift Feature ◽  
Matching Pair

The image stitching method is widely used into the suspect's footprint information extraction. In order to improve the image detail and the matching precision, the Footprint map image stitching method which is based on the wavelet transform and the SIFT feature matching is put forward. The wavelet transform in this method is perform based on the pretreatment of image, move the low frequency wavelet coefficient to zero, adjusting thresholds of the high frequency wavelet coefficient and inverse transformation, then, use the SIFT to extract and match the key-points of the processed images. For the error matching pair of coarse match, you can use the RANSAC to filter them out. This article demonstrates its advantage through to the original image splicing comparisons. The experimental results show that the method display more clear detail and the precision of matching than the original method.

Infrared Image and Visible Image Fusion Based on Wavelet Transform

2013 ◽  
Vol 756-759 ◽  
pp. 2850-2856 ◽  
Author(s):  
Ze Hua Zhou ◽  
Min Tan
Keyword(s):  
Wavelet Transform ◽  
Image Fusion ◽  
Infrared Image ◽  
Low Frequency ◽  
Visible Image ◽  
Detection And Identification ◽  
Single Sensor ◽  
Physical Phenomena ◽  
Image Fusion Method

The same scene, the infrared image and visible image fusion can concurrently take advantage of the original image information can overcome the limitations and differences of a single sensor image in terms of geometric, spectral and spatial resolution, to improve the quality of the image , which help to locate, identify and explain the physical phenomena and events. Put forward a kind of image fusion method based on wavelet transform. And for the wavelet decomposition of the frequency domain, respectively, discussed the principles of select high-frequency coefficients and low frequency coefficients, highlight the contours of parts and the weakening of the details section, fusion, image fusion has the characteristics of two or multiple images, more people or the visual characteristics of the machine, the image for further analysis and understanding, detection and identification or tracking of the target image.

Sign in / Sign up

Export Citation Format

Share Document