Local Feature Points Research Articles

Accurate and robust estimation of camera parameters using RANSAC

Camera calibration plays an important role in the field of machine vision applications. The popularly used calibration approach based on 2D planar target sometimes fails to give reliable and accurate results due to the inaccurate or incorrect localization of feature points. To solve this problem, an accurate and robust estimation method for camera parameters based on RANSAC algorithm is proposed to detect the unreliability and provide the corresponding solutions. Through this method, most of the outliers are removed and the calibration errors that are the main factors influencing measurement accuracy are reduced. Both simulative and real experiments have been carried out to evaluate the performance of the proposed method and the results show that the proposed method is robust under large noise condition and quite efficient to improve the calibration accuracy compared with the original state.

Active Shape and Active Apparent Hybrid Model of Human Facial Feature Points Localization Algorithm

An improved active shape model (ASM) and active appearance model (AAM) based new method is proposed, this method will use two-dimensional local gray information to update the feature point position when using ASM to locate the face contour feature points. As to the internal features point location, it establishes facial organs independent AAM model. At the same time, it uses different measure functions to judge the convergence of search algorithm. The experimental results show that the new algorithm greatly improved the localization accuracy of facial feature points.

ORIENTATION BASED BUILDING OUTLINE EXTRACTION IN AERIAL IMAGES

Abstract. The goal of this paper is to extract automatically the building contours regardless of shape. By extracting these contours, detection results will be more accurate, giving useful information about urban area, which is important for many tasks, like map updating and disaster management. First, we extract local feature points from the image, based on a modification of Harris detector's saliency function, which can represent urban area and building effectively. This point set is then used to define the main orientation of the buildings, which characterizes well an urban region and helps to define directions, where object contours have to be searched. Second, we applied shearlet approach to extract edges in the defined directions. This results in an edge map, which helps us to determine point subsets belonging to the same building. Convex hulls of the point subsets is used for contour initialization, then region based Chan-Vese Active Contour method is applied to extract the accurate building outlines.

LF-EME: Local features with elastic manifold embedding for human action recognition

Human action recognition has been an active topic in computer vision. Currently, most of the approaches to this problem can be categorized into two classes. One is based on local features, and the other is based on global features. Meanwhile, manifold learning has become successful in many problems in computer vision, but because of the high variability of human body, the application of manifold learning to human action recognition is limited. We propose a framework based on Elastic Manifold Embedding (EME), a new sparse manifold learning algorithm, together with local interest point features to handle human action recognition. The result of the new framework is very promising in comparison with state-of-the-art methods.

Predictive control of nonlinear visual servoing systems using image moments

The visual servoing architectures with an eye-in-hand configuration, composed of a 6-d.o.f. manipulator robot and a video camera, are categorised as high-coupled and nonlinear systems. For this reason, in order to have a stable and convergent behaviour, advanced techniques like predictive control are needed. These types of techniques can solve the potential problems that can appear in classical (image-based visual servoing) while different type of constraints for visual servoing system can be fulfilled easier. Three different visual predictive control architectures based on image moments are proposed. The model for the image moments predictor used in the first architecture is an extension of a local model point features predictor. The second predictive approach uses an image moment-based predictor to describe the future behaviour of the system over a certain camera velocity scenario directly in the image moments space. In order to have a faster convergence and in the same time a desired behaviour of the servoing system, a reference trajectory is employed. This represents the main contribution of the last predictive control technique proposed in this study. A Matlab simulator was developed and all the visual servoing architectures described in the study were implemented. The simulation results showed that all the proposed visual predictive control approaches are stable and convergent when dealing with visual servoing tasks, but the best performances were assigned to the reference trajectory-based image moments architecture.

Illumination-Invariant Localization Based on Upward Looking Scenes for Low-Cost Indoor Robots

This paper presents a novel method for localizing a robot using an upward-looking camera under significant illumination changes. The proposed method uses two different types of visual features that are efficiently and robustly detected in upward looking indoor scenes: the orthogonal lines and local descriptor-based point features. The dominant orientation of distinguishable orthogonal lines in indoor environments is utilized to correct the robot orientation. This approach is very simple and powerful since the feature matching problem does not need to be considered when computing the dominant orientation of orthogonal lines. The local descriptors are used to effectively match local point features between two images. We present how to correct the robot location with a few noisy matches caused by a significant change of the illumination. The proposed localization method was implemented as an embedded system with an ARM11 processor on a home cleaning robot. Our experimental results show that the proposed method works well in indoor environments with various illumination changes.

Geometrically Invariant Watermarking Scheme Based on Local Feature Points

Based on local invariant feature points and cross ratio principle, this paper presents a feature-point-based image watermarking scheme. It is robust to geometric attacks and some signal processes. It extracts local invariant feature points from the image using the improved scale invariant feature transform algorithm. Utilizing these points as vertexes it constructs some quadrilaterals to be as local feature regions. Watermark is inserted these local feature regions repeatedly. In order to get stable local regions it adjusts the number and distribution of extracted feature points. In every chosen local feature region it decides locations to embed watermark bits based on the cross ratio of four collinear points, the cross ratio is invariant to projective transformation. Watermark bits are embedded by quantization modulation, in which the quantization step value is computed with the given PSNR. Experimental results show that the proposed method can strongly fight more geometrical attacks and the compound attacks of geometrical ones.

Perceptual Image Hashing Based on Shape Contexts and Local Feature Points

Local feature points have been widely investigated in solving problems in computer vision, such as robust matching and object detection. However, its investigation in the area of image hashing is still limited. In this paper, we propose a novel shape-contexts-based image hashing approach using robust local feature points. The contributions are twofold: 1) The robust SIFT-Harris detector is proposed to select the most stable SIFT keypoints under various content-preserving distortions. 2) Compact and robust image hashes are generated by embedding the detected local features into shape-contexts-based descriptors. Experimental results show that the proposed image hashing is robust to a wide range of distortions and attacks, due to the benefits of robust salient keypoints detection and the shape-contexts-based feature descriptors. When compared with the current state-of-the-art schemes, the proposed scheme yields better identification performances under geometric attacks such as rotation attacks and brightness changes, and provides comparable performances under classical distortions such as additive noise, blurring, and compression. Also, we demonstrate that the proposed approach could be applied for image tampering detection.

Pose and Expression Independent Facial Landmark Localization Using Dense-SURF and the Hausdorff Distance

We present an approach to automatic localization of facial feature points which deals with pose, expression, and identity variations combining 3D shape models with local image patch classification. The latter is performed by means of densely extracted SURF-like features, which we call DU-SURF, while the former is based on a multiclass version of the Hausdorff distance to address local classification errors and nonvisible points. The final system is able to localize facial points in real-world scenarios, dealing with out of plane head rotations, expression changes, and different lighting conditions. Extensive experimentation with the proposed method has been carried out showing the superiority of our approach with respect to other state-of-the-art systems. Finally, DU-SURF features have been compared with other modern features and we experimentally demonstrate their competitive classification accuracy and computational efficiency.

Human fringe skeleton extraction by an improved Hopfield neural network with direction features

The fringe skeleton is a useful means of shape description to 3D human bodies. A new method is proposed to extract human fringe skeleton by matching feature points of torso and tracking on direction of limb. We primarily extract the depth direction feature based on coronal plane of 3D human model. By introducing three transverse cross sections on torso and the concept of triangle pairs, the extraction result of coronal plane and its depth direction feature will not be affected by shading and interference of limb on different postures. With the local direction feature, an improved Hopfield neural network (IMHNN) is used to locate feature points of any position by matching feature points of template model and the personalized target model. Subsequently feature points on limb are extracted by limb direction feature using shoulder and hip feature points. Finally, the fringe skeleton of human model is obtained by the combination of feature points on torso and limb. Experiments show that the proposed algorithm has lower computation time and better extracted results compared with the existing fringe skeleton extraction algorithms.

The Fast Image Matching Algorithm for Building Photogrammetry

SIFT is the most common algorithm for the image local feature points matching. The excellency of it is not only good spatial scale invariance, but also more accurate and faster than other algorithm. However, the SIFT feature points do not reflect the geometric features of objects, so, when dealing with the building images, these points are not available in most cases, and the extraction process is complicated. Therefore, this paper presents a new algorithm that combines the Harris corner detector and SIFT operator. This new algorithm not only can enhance the efficiency of image matching, and make accurate information on the building corner, but also provide good reference information for modeling. Experiments show that the extract feature points of this algorithm can be applied to the three-dimensional reconstruction of large buildings.

Unorganized 3-D Points Data Registration with Local Geometric Feature Based on CUDA

Automatic Automatic 3-D unorganized point data registration technique,which maps 3-D datas measured from multiple viewpoints into a common coordinate space, is a key technique for reverse engineering.In order to improve data matching speed, a parallel detecting algorithm with local geometric feature based on CUDA architechure was proposed.Firstly,local geometric feature points were extracted from original data sets,and then the correspondence between them are computed, at last this registration algorithm was implemented in parallel pattern on CUDA.The comparison experiments show that this algorithm is efficient and robust against noise.

Identifying biological landmarks using a novel cell measuring image analysis tool: Cell-o-Tape.

BackgroundThe ability to quantify the geometry of plant organs at the cellular scale can provide novel insights into their structural organization. Hitherto manual methods of measurement provide only very low throughput and subjective solutions, and often quantitative measurements are neglected in favour of a simple cell count.ResultsWe present a tool to count and measure individual neighbouring cells along a defined file in confocal laser scanning microscope images. The tool allows the user to extract this generic information in a flexible and intuitive manner, and builds on the raw data to detect a significant change in cell length along the file. This facility can be used, for example, to provide an estimate of the position of transition into the elongation zone of an Arabidopsis root, traditionally a location sensitive to the subjectivity of the experimenter.ConclusionsCell-o-tape is shown to locate cell walls with a high degree of accuracy and estimate the location of the transition feature point in good agreement with human experts. The tool is an open source ImageJ/Fiji macro and is available online.

A METHOD FOR FACE RECOGNITION USING IMAGE REGISTRATION

This paper presents a technique for face recognition that is based on image registration. The face recognition technique consists of three parts: a training part, an image registration part and a post-processing part. The image registration technique is based on finding a set of feature points in the two images and using these feature points for registration. This is done in four steps. In the first, images are filtered with the Mexican-hat wavelet to obtain the feature point locations. In the second, the Zernike moments of neighborhoods around the feature points are calculated and compared in the third step to establish correspondence between feature points in the two images. In the fourth, the transformation parameters between images are obtained using an iterative least squares technique to eliminate outliers.1,2 During training, a set of images are chosen as the training images and the Zernike moments for the feature points of the training images are obtained and stored. The choice of training images depends on the changes of poses and illumination that are expected. In the registration part, the transformation parameters to register the training images with the images under consideration are obtained. In the post-processing, these transformation parameters are used to determine whether a valid match is found or not. The performance of the proposed method is evaluated using various face databases3–5 and it is compared with the performance of existing techniques. Results indicate that the proposed technique gives excellent results for face recognition in conditions of varying pose, illumination, background and scale.

Action unit classification using active appearance models and conditional random fields

In this paper, we investigate to what extent modern computer vision and machine learning techniques can assist social psychology research by automatically recognizing facial expressions. To this end, we develop a system that automatically recognizes the action units defined in the facial action coding system (FACS). The system uses a sophisticated deformable template, which is known as the active appearance model, to model the appearance of faces. The model is used to identify the location of facial feature points, as well as to extract features from the face that are indicative of the action unit states. The detection of the presence of action units is performed by a time series classification model, the linear-chain conditional random field. We evaluate the performance of our system in experiments on a large data set of videos with posed and natural facial expressions. In the experiments, we compare the action units detected by our approach with annotations made by human FACS annotators. Our results show that the agreement between the system and human FACS annotators is higher than 90% and underlines the potential of modern computer vision and machine learning techniques to social psychology research. We conclude with some suggestions on how systems like ours can play an important role in research on social signals.

Enhance ASM Based on DCT-SVM for Facial Feature Points Localization

Focused on the facial feature points localization, the enhance ASM algorithm based on modeling texture by the DCT-SVM is proposed. First, the statistical shape model is built. Then, some key feature points are selected and their texture models are built by the DCT-SVM. In the subsequent searching, the feature points are divided into two classes based on their reliability gained by DCT-SVM detector, by combining the reliable feature points to the shape constraint, the original shape can finally match to the target face. Experiments show the algorithm is robust to the expressions change and can better locate the features than the traditional ASM.

Iris Recognition Using Possibilistic Fuzzy Matching on Local Features

In this paper, we propose a novel possibilistic fuzzy matching strategy with invariant properties, which can provide a robust and effective matching scheme for two sets of iris feature points. In addition, the nonlinear normalization model is adopted to provide more accurate position before matching. Moreover, an effective iris segmentation method is proposed to refine the detected inner and outer boundaries to smooth curves. For feature extraction, the Gabor filters are adopted to detect the local feature points from the segmented iris image in the Cartesian coordinate system and to generate a rotation-invariant descriptor for each detected point. After that, the proposed matching algorithm is used to compute a similarity score for two sets of feature points from a pair of iris images. The experimental results show that the performance of our system is better than those of the systems based on the local features and is comparable to those of the typical systems.

Myocardial ischemia analysis based on electrocardiogram QRS complex in time domain

Objective ST-T complex change, which represents the ventricle repolarization phase, is the main clinical indicator in detecting myocardial ischemia (MI) based on electrocardiogram (ECG) signals.However, its feature point location is not accurate due to interferences. In this paper, a new approach about myocardial ischemia analysis was proposed based on QRS complex. Methods QRS complex, representing the ventricle depolarization process, was used to analyze myocardial ischemia, and some parameters were extracted synthetically in time domain. Then they were used for statistical analysis of myocardial ischemia states and non-myocardial ischemia states. Results Five parameters had significant differences after verification of Non-MI signals in MIT-BIH database and MI signals in long-term ST database (LTST) and they were: QRS upward and downward slopes, transient heart rate, R angle and Q angle in a triangle QRS. Conclusion Five parameters extracted from QRS complex had significant differences. The proposed method provides an important basis for myocardial ischemia detection. Key words: Electrocardiogram; Myocardial ischemia; QRS complex; Upward slope

Robust Low Complexity Corner Detector

Corner feature point detection with both the high-speed and high-quality is still very demanding for many real-time computer vision applications. The Harris and Kanade-Lucas-Tomasi (KLT) are widely adopted good quality corner feature point detection algorithms due to their invariance to rotation, noise, illumination, and limited view point change. Although they are widely adopted corner feature point detectors, their applications are rather limited because of their inability to achieve real-time performance due to their high complexity. In this paper, we redesigned Harris and KLT algorithms to reduce their complexity in each stage of the algorithm: Gaussian derivative, cornerness response, and non-maximum suppression (NMS). The complexity of the Gaussian derivative and cornerness stage is reduced by using an integral image. In NMS stage, we replaced a highly complex sorting and NMS by the efficient NMS followed by sorting the result. The detected feature points are further interpolated for sub-pixel accuracy of the feature point location. Our experimental results on publicly available evaluation data-sets for the feature point detectors show that our low complexity corner detector is both very fast and similar in feature point detection quality compared to the original algorithm. We achieve a complexity reduction by a factor of 9.8 and attain 50 f/s processing speed for images of size 640×480 on a commodity central processing unit with 2.53 GHz and 3 GB random access memory.

Robust video watermarking based on affine invariant regions in the compressed domain

This paper proposes a novel robust video watermarking scheme based on local affine invariant features in the compressed domain. This scheme is resilient to geometric distortions and quite suitable for DCT-encoded compressed video data because it performs directly in the block DCTs domain. In order to synchronize the watermark, we use local invariant feature points obtained through the Harris-Affine detector which is invariant to affine distortions. To decode the frames from DCT domain to the spatial domain as fast as possible, a fast inter-transformation between block DCTs and sub-block DCTs is employed and down-sampling frames in the spatial domain are obtained by replacing each sub-blocks DCT of 2×2 pixels with half of the corresponding DC coefficient. The above-mentioned strategy can significantly save computational cost in comparison with the conventional method which accomplishes the same task via inverse DCT (IDCT). The watermark detection is performed in spatial domain along with the decoded video playing. So it is not sensitive to the video format conversion. Experimental results demonstrate that the proposed scheme is transparent and robust to signal-processing attacks, geometric distortions including rotation, scaling, aspect ratio changes, linear geometric transforms, cropping and combinations of several attacks, frame dropping, and frame rate conversion.