Local Invariant Features Research Articles

Improved WαSH Feature Matching Based on 2D-DWT for Stereo Remote Sensing Images.

Image matching is an outstanding issue because of the existing of geometric and radiometric distortion in stereo remote sensing images. Weighted α-shape (WαSH) local invariant features are tolerant to image rotation, scale change, affine deformation, illumination change, and blurring. However, since the number of WαSH features is small, it is difficult to get enough matches to estimate the satisfactory homography matrix or fundamental matrix. In addition, the WαSH detector is extremely sensitive to image noise because it is built on sampled edges. Considering the shortcomings of the WαSH detector, this paper improves the WαSH feature matching method based on the 2D discrete wavelet transform (2D-DWT). The method firstly performs 2D-DWT on the image, and then detects WαSH features on the transformed images. According to the methods of descriptor construction for WαSH features, three matching methods on the basis of wavelet transform WαSH features (WWF), improved wavelet transform WαSH features (IWWF), and layered IWWF (LIWWF) are distinguished with respect to the character of the sub-images. The experimental results on the dataset containing affine distortion, scale distortion, illumination change, and noise images, showed that the proposed methods acquired more matches and better stableness than WαSH. Experimentation on remote sensing images with less affine distortion and slight noise showed that the proposed methods obtained the correct matching rate greater than 90%. For images containing severe distortion, KAZE obtained a 35.71% correct matching rate, which is unacceptable for calculating the homography matrix, while IWWF achieved a 71.42% correct matching rate. IWWF was the only method that achieved the correct matching rate of no less than 50% for all four test stereo remote sensing image pairs and was the most stable compared to MSER, DWT-MSER, WαSH, DWT-WαSH, KAZE, WWF, and LIWWF.

A spatial-spectral SIFT for hyperspectral image matching and classification

The scale-invariant feature transform (SIFT) is known as one of the most robust local invariant feature and is widely applied to image matching and classification. However, There is few studies on SIFT for hyperspectral image (HSI). Hyperspectral image (HSI) embraces the spectral information reflecting the material radiation property and the geometrical relationship of the objects. Thus, HSI provides much more information than gray and color image. Therefore, this paper puts forward a spatial-spectral SIFT for HSI matching and classification by using the geometric algebra as its mathematic tool. It extracts and describes the spatial-spectral SIFT feature in the spatial-spectral domain to exploit both the spectral and spatial information of HSI. Firstly, a spatial-spectral unified model of spectral value and gradient change (UMSGC for short) is built to analyze spectral and spatial information for HSI synthetically. Secondly, the scale space for HSI based on UMSGC is designed. Finally, both the new detector and descriptor of the spatial-spectral SIFT for HSI that comprehensively consider spectral and spatial information are proposed. The experimental results show that the proposed algorithm demonstrates excellent performance in HSI matching and classification.

Effective Characterization of Relief Patterns

AbstractIn this paper, we address the problem of characterizing relief patterns over surface meshes independently on the underlying shape. We propose to tackle the problem by estimating local invariant features and encoding them using the Improved Fisher Vector technique, testing both features estimated on 3D meshes and local descriptors estimated on raster images created by encoding local surface properties (e.g. mean curvature) over a surface parametrization. We compare the robustness of the obtained descriptors against noise and surface bending and evaluate retrieval performances on a specific benchmark proposed in a track of the Eurographics Shape REtrieval Contest 2017. Results show that, with the proposed framework, it is possible to obtain retrieval results largely improving the state of the art and that the image‐based approach is still effective when the underlying surface is heavily deformed.

A local phase based invariant feature for remote sensing image matching

Local invariant features from computer vision community have recently been widely applied to the matching of remote sensing images. However, these features are mainly designed to handle geometric distortions, and are sensitive to complex radiometric differences between multisensor images. To address this issue, this paper proposes an effective local invariant feature that is sufficiently robust to both geometric and radiometric changes. The proposed feature is built based on the phase congruency model that is invariant to illumination and contrast variation. It consists of a feature detector named MMPC-Lap and a feature descriptor named local histogram of orientated phase congruency (LHOPC). MMPC-Lap is constructed by using the minimum moment of phase congruency for feature detection with an automatic scale location technique, which is used to detect stable keypoints in image scale space. Subsequently, LHOPC derives the feature descriptor for a keypoint by utilizing an extended phase congruency feature with an advanced descriptor configuration. Finally, correspondences are achieved by evaluating the similarity of the feature descriptors. The proposed MMPC-Lap and LHOPC have been evaluated under different imaging conditions (spectral, temporal, and scale changes). The results obtained on a variety of remote sensing images demonstrate its excellent performance with respect to the state-of-the-art local invariant features, especially for cases where there are complex radiometric differences.

Remote sensing image retrieval based on sparse local invariant features

Remote sensing image retrieval based on sparse local invariant features

Extreme-constrained spatial-spectral corner detector for image-level hyperspectral image classification

As one type of local invariant feature, corner feature plays an important role in diverse applications such as: video mining, target detection, image classification, image retrieval, and image matching, etc. However, there are few studies on corner feature for hyperspectral image (HSI). Therefore, this paper proposes a novel corner feature for HSI named extreme-constrained spatial-spectral corner (ECSSC for short) and its corresponding detector. The definition of ECSSC is developed based on the definition of spectral-spatial interest point and the characteristic of HSI. Based on this definition, the detector of ECSSC is put forward and introduced in detail. Then, as an important application of ECSSC, an efficient framework for image-level HSI classification is designed based on ECSSC and parallel computation. The experimental results show that the proposed algorithm can detect abundant corner features with high repeatability rate from HSI and the accuracy of image-level HSI based on ECSSC is dramatically higher than that of the state of the art.

Recognition and retrieval of objects in diverse applications

This work proposes and evaluates object description and retrieval techniques in different real applications. It addresses the classification of boar spermatozoa according to acrosome integrity using several methods based on invariant local features. In addition, it provides two new methods for insert localisation and an automatic solution for the recognition of broken inserts in edge profile milling heads that can be set up in-process without delaying any machining operations. Finally, it evaluates different clusterings of keypoints for object retrieval and proposes a new descriptor, named colour COSFIRE, in the scope of the European project Advisory System Against Sexual Exploitation of Children.

Indexing near-duplicate images in web search using minhash algorithm

Near-duplicate images cause problems of redundancy and copyright infringement in large image collections. The trouble is minor in the web, where appropriation of images without acknowledgment of source is prevalent. Near duplicates can be exact copies or else differ slightly in their visual content. Near-duplicate detection has received substantial attention over the past few years due to the applications in copyright enforcement, organizing large size image databases, increasing focus on image search, redundancy elimination of logos, managing storage space by removing duplication, etc. In the paper, a method has been proposed for Indexing Near-Duplicate Images in the Web Search. First image enhancement is done in user query image then features are extracted based on SURF (Speeded up Robust Features) that is to extract the local invariant features of each image. After this similarity measure is calculated among the feature extracted images using min-hash algorithm. Finally, Locality Sensitive Hashing (LSH) is used for indexing near duplicate images based on user query. We demonstrate that our indexing approach is highly effective for collections of up to a few hundred thousand images.

A Novel SURF Based on a Unified Model of Appearance and Motion-Variation

The speeded-up robust features (SURFs) algorithm is the best and most efficient local invariant feature algorithm for application to 2-D images and is widely applied in the fields of 2-D image processing and computer vision. Compared to 2-D images, a video has motion information in addition to its appearance information. Here, to make full use of the video appearance and motion information, we use geometric algebra as the mathematical calculation and analysis framework to obtain the embedded appearance and motion information on a local area of a video. In our proposed model of appearance and motion variation (UMAMV), we developed SURF feature detection and description algorithms operating on the spatio-temporal domain with video appearance and motion information. First of all, a model of appearance and motion variation, which contains video appearance and motion information in the framework of geometric algebra, is proposed. Then, based on this model, we propose a novel detection algorithm, theUMAMV-SURF detector, which mainly contains Hessian matrix construction, Hessian matrix determinant approximation calculation, and non-maximal suppression determination feature points as its key steps. Then, we introduce the UMAMV-SURF description algorithm, which mainly includes determining the dominant orientation of UMAMV-SURF feature points and generating the UMAMV-SURF feature descriptors. Finally, by experimenting with the Weizman and UCF101 datasets, the experimental results show that the proposed UMAMV-SURF algorithm can detect those SURF feature points which can have unique appearance information in the spatial domain and reflect motion change in the temporal domain. Moreover, it offers a higher accuracy than other spatio-temporal interest point algorithms in human behavior recognition of video footage.

Performance Characterization of Image Feature Detectors in Relation to the Scene Content Utilizing a Large Image Database

Selecting the most suitable local invariant feature detector for a particular application has rendered the task of evaluating feature detectors a critical issue in vision research. Although the literature, offers a variety of comparison works focusing on performance evaluation of image feature detectors under several types of image transformations, the influence of the scene content on the performance of local feature detectors has received little attention so far. This paper, aims to bridge this gap with a new framework for determining the type of scenes which maximize and minimize the performance of detectors in terms of repeatability rate. The results are presented for several state-of-the-art feature detectors that have been obtained using a large image database of 20482 images under JPEG compression, uniform light and blur changes with 539 different scenes captured from real-world scenarios. These results provide new insights into the behavior of feature detectors.

An efficient categorization of liver cirrhosis using convolution neural networks for health informatics

Accurate categorization of cirrhosis liver image in ultrasound modality is of great importance in medical diagnosis and treatment. Health informatics is important as it provides quick predictions on diseases based on nearness of symptoms. Modelling solutions for the same on cloud using deep learning is the motivation of this paper. Here, we propose a deep learning model associated with correlation based feature selection method for cirrhosis image classification. We compare the results with three other conventional classifiers algorithms to improve the better classification accuracy. First in pre-processing stage, noises are eliminated from pathological scan images by using modified laplacian pyramid non-linear diffusion filter. From the pre-processed scan images, each cirrhosis region is obtained under the guidance of radiology or physicians. Then, after extracting the complete features of each patch by gray level, local binary pattern and scale invariant feature, a feature selection technique is applied to choice the predominant texture features for each classifier. Finally a convolution neural network is implemented to improve the performance of classifiers in terms of sensitivity, specificity and accuracy. Convolution neural network algorithm with two hidden layers gives more accuracy in classifying cirrhosis image with 98% sensitivity. Experiments are carried out with 990 cirrhosis image patches which demonstrates that our proposed deep learning classifier perform 100% well than original classifiers in terms of accuracy.

Palm vein recognition method based on fusion of local Gabor histograms

Gabor features have been shown to be effective for palm vein recognition. This paper presents a novel feature representation method, implementing the fusion of local Gabor histograms (FLGH), in order to improve the accuracy of palm vein recognition systems. A new local descriptor called local Gabor principal differences patterns (LGPDP) encodes the Gabor magnitude using the local maximum difference (LMD) operator. The corresponding Gabor phase patterns are encoded by local Gabor exclusive OR (XOR) patterns (LGXP). Fisher's linear discriminant (FLD) method is then implemented to reduce the dimensionality of the feature representation. Low-dimensional Gabor magnitude and phase feature vectors are finally fused to enhance accuracy. Experimental results from Institute of Automation, Chinese Academy of sciences (CASIA) database show that the proposed FLGH method achieves better performance by utilizing score-level fusion. The equal error rate (EER) is 0.08%, which outperforms other conventional palm vein recognition methods (EER range from 2.87% to 0.16%), e.g., the Laplacian palm, minutiae feature, Hessian phase, Eigenvein, local invariant features, mutual foreground local binary patterns (LBP), and multi-sampling feature fusion methods.

A unified model of appearance and motion of video and its application in STIP detection

Spatio-temporal interest points (STIPs) are local invariant features of a video and are both distinctive and descriptive and therefore can be applied in action recognition; however, most existing STIP detectors extend spatial descriptions by adding a temporal component for the appearance description, which separate spatio-temporal domain correlations in the spatio-temporal domain and only implicitly capture motion information. Therefore, by regarding the video as a 3-d structure, this research aims to develop a novel STIP detector which synthetically exploits appearance and motion information using Clifford algebra. This study firstly establishes a general Clifford algebra model for video and then builds the unified model of appearance and motion (UMAM) based thereon to synthetically analyse appearance and motion information. Subsequently, in the spirit of the well-known Harris 3-d detector, a UMAM-based spatio-temporal Harris corner detector (UMAM-Harris) for videos is developed. The experimental results indicate that the UMAM-Harris detector proposed in this study extracts the UMAM-Harris corners that contain distinctive features in the spatial domain and reflect substantial motion in the time domain, and it offers a better performance than the traditional STIP detection algorithms used in video action recognition.

Object tracking by transitive learning using perspective transformation with asymptotic stability

Object tracking is a core subject in computer vision and has significant meaning in both theory and practice. We propose a tracking method in which a robust discriminative classifier is built based on both object and context information. In this method, we consider multiple frames of local invariant features on and around the object and construct the object template and context template. To overcome the limitation of the invariant representations, we also design a nonparametric learning algorithm using transitive matching perspective transformation. This learning algorithm can keep adding object appearance and can avoid improper updating when occlusions appear. We also analyze the asymptotic stability of our method and prove its drift-free capability in long-term tracking. Extensive experiments using challenging publicly available video sequences that cover most of the critical conditions in tracking demonstrate the enhanced strength and robustness of our method.

A Modified Image Registration Algorithm Based on Color Invariance and RANSAC

Image category recognition and feature matching is an important access to visual information on the level of objects and scenes. So far, most works concerning local feature matching are limited in gray images. Designed for gray images, SIFT has been found to be a robust local invariant feature descriptor. However, much information is ignored since the color is discarded in traditional SIFT, and some mismatching problems happen due to this approach. To achieve a more accurate feature extraction and keypoints matching, more advanced method is desperately needed. In this paper, we present a group of color descriptors, to increase the performance traditional SIFT does in some occasions like outdoor scenes. Besides, a RANSAC algorithm is also analyzed and applied to our research, to leave out those obvious mismatching. RANSAC works well for images with many self-similar structures, which decreases mismatches significantly. Then we improved the method so that it converges in fewer iterations.

Distortion calibrating method of measuring rail profile based on local affine invariant feature descriptor

Measuring rail profile in the presence of multiple degrees of freedom vibration is a very challenging task. This paper presents a novel method based on the local affine invariant feature descriptor to calibrate distorted profiles, which are obtained by traditional rail measurement system. It has three major modules: local affine invariant (LAI) feature descriptor, affine transformation estimation and parameters refinement. LAI feature descriptor is based on the affine geometry invariant and generated by calculating the proportions of different areas. Using the proposed LAI descriptor, we implement a three-stage profile calibration including matching, estimation, and refinement based on grouping and fast iterative closest point (FICP) algorithm. The performance of proposed LAI descriptor and calibrating method is tested by performing extensive experiments. The experimental results show that our LAI descriptor is highly descriptive and robust with respect to varying resolution and noise, and the LAI descriptor based calibration is effective and repeatable.

Fusing Local and Global Features for High-Resolution Scene Classification

In this paper, a fused global saliency-based multiscale multiresolution multistructure local binary pattern (salM 3LBP) feature and local codebookless model (CLM) feature is proposed for high-resolution image scene classification. First, two different but complementary types of descriptors (pixel intensities and differences) are developed to extract global features, characterizing the dominant spatial features in multiple scale, multiple resolution, and multiple structure manner. The micro/macrostructure information and rotation invariance are guaranteed in the global feature extraction process. For dense local feature extraction, CLM is utilized to model local enrichment scale invariant feature transform descriptor and dimension reduction is conducted via joint low-rank learning with support vector machine. Finally, a fused feature representation between salM3LBP and CLM as the scene descriptor to train a kernel-based extreme learning machine for scene classification is presented. The proposed approach is extensively evaluated on three challenging benchmark scene datasets (the 21-class land-use scene, 19-class satellite scene, and a newly available 30-class aerial scene), and the experimental results show that the proposed approach leads to superior classification performance compared with the state-of-the-art classification methods.

RGBD Human Action Recognition using Multi-Features Combination and K-Nearest Neighbors Classification

In this paper, we present a novel system to analyze human body motions for action recognition task from two sets of features using RGBD videos. The Bag-of-Features approach is used for recognizing human action by extracting local spatialtemporal features and shape invariant features from all video frames. These feature vectors are computed in four steps: Firstly, detecting all interest keypoints from RGB video frames using Speed-Up Robust Features and filters motion points using Motion History Image and Optical Flow, then aligned these motion points to the depth frame sequences. Secondly, using a Histogram of orientation gradient descriptor for computing the features vector around these points from both RGB and depth channels, then combined these feature values in one RGBD feature vector. Thirdly, computing Hu-Moment shape features from RGBD frames; fourthly, combining the HOG features with Hu-moments features in one feature vector for each video action. Finally, the k-means clustering and the multi-class K-Nearest Neighbor is used for the classification task. This system is invariant to scale, rotation, translation, and illumination. All tested, are utilized on a dataset that is available to the public and used often in the community. By using this new feature combination method improves performance on actions with low movement and reach recognition rates superior to other publications of the dataset.

Contact lens detection for iris spoofing countermeasure

The development of biometric authentication system should be followed by strengthening to spoofing attempts. Among various identifiers, iris has aroused many attentions due to its uniqueness and stability. Nevertheless, the use of iris for biometric authentication is accompanied by spoofing risk, for example using contact lens. In order to handle the spoofing attempts, its detection is an inevitable part of a recognition system, to reduce the risk of forging system. Cosmetic contact lens is one of most common spoofing materials which is hard to be detected. In this study, weighted local binary pattern (w-LBP) and simplified scale invariant feature transform (SIFT) descriptors were used to extract the feature of the iris, in which segmented using gradient magnitude and Fourier descriptor. Simplified SIFT descriptor is extracted at each pixel of iris image and being used to rank the local binary pattern (LBP) sequence of encoding. The features were then presented to support vector machine (SVM) classifier, for positive vs. negative classification. Positive class means that contact lens was used by a person, and vice versa. The experimental results showed that combining SIFT and w-LBP as features for SVM yielded an accuracy of 84%.

A new modified panoramic UAV image stitching model based on the GA-SIFT and adaptive threshold method

This paper presents panoramic unmanned aerial vehicle (UAV) image stitching techniques based on an optimal Scale Invariant Feature Transform (SIFT) method. The image stitching representation associates a transformation matrix with each input image. In this study, we formulate stitching as a multi-image matching problem, and use invariant local features to find matches between the images. An improved Geometric Algebra (GA-SIFT) algorithm is proposed to realize fast feature extraction and feature matching work for the scanned images. The proposed GA-SIFT method can locate more feature points with greater accurately than the traditional SIFT method. The adaptive threshold value method proposed solves the limitation problem of high computation load and high cost of stitching time by greater feature points extraction and stitching work. The modified random sample consensus method is proposed to estimate the image transformation parameters and to determine the solution with the best consensus for the data. The experimental results demonstrate that the proposed image stitching method greatly increases the speed of the image alignment process and produces a satisfactory image stitching result. The proposed image stitching model for aerial images has good distinctiveness and robustness, and can save considerable time for large UAV image stitching.