Object Localization Method Research Articles

Fast Fourier-based simulation of off-resonance artifacts in steady-state gradient echo MRI applied to metal object localization.

PurposeTo accelerate simulation of off‐resonance artifacts in steady‐state gradient echo MRI by using fast Fourier transforms and demonstrate its applicability to metal object localization.Theory and MethodsBy exploiting the repetitive nature of steady‐state pulse sequences it is possible to use fast Fourier transforms to calculate the MR signal. Based on this principle, a method for fast simulation of off‐resonance artifacts was designed. The method was validated against Bloch simulations and MRI scans. Its clinical relevance was demonstrated by employing it for template matching‐based metal object localization, as applied to a titanium cylinder, an oxidized zirconium knee implant, and gold fiducials.ResultsThe fast simulations were accurate compared with actual MRI scans of the objects. The differences between the fast simulations and Bloch simulations were minor, while the acceleration scaled linearly with the number of phase‐encoding lines. The object localization method accurately localized the various metal objects.ConclusionThe proposed simulation methodology provided accurate 3D simulations of off‐resonance artifacts with a lower computational complexity than Bloch simulations. The speed of the simulations opens up possibilities in image reconstructions involving off‐resonance phenomena that were previously infeasible due to computational limitations, as demonstrated for metal object localization. Magn Reson Med 78:2035–2041, 2017. © 2016 The Authors Magnetic Resonance in Medicine published by Wiley Periodicals, Inc. on behalf of International Society for Magnetic Resonance in Medicine. This is an open access article under the terms of the Creative Commons Attribution NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

A non-cooperative target grasping position prediction model for tethered space robot

How to identify proper grasping positions on non-cooperative targets is an essential technology for tethered space robots when implementing on-orbit service. To localize unknown targets, most template-based methods require a large set of well-engineered templates. Sliding them over the image to decide potential target positions is exhaustive and inefficient. To solve this rather complex problem, we propose a novel object localization method by predicting object regions before extracting them. It can reduce the search area of targets remarkably and runs fast. Firstly, the features of histogram of oriented gradients are modified to be more discriminative. Then the cascaded support vector machine is used to select better proposals over scales and aspect ratios. We also integrate the pre-processing procedure to highlight active target features. Further experiments demonstrate that our method improves the detection rate in VOC 2007 favorably and performs well in satellite brackets localization.

Accurate Segmentation for Infrared Flying Bird Tracking

Bird strikes present a huge risk for air vehicles, especially since traditional airport bird surveillance is mainly dependent on inefficient human observation. For improving the effectiveness and efficiency of bird monitoring, computer vision techniques have been proposed to detect birds, determine bird flying trajectories, and predict aircraft takeoff delays. Flying bird with a huge deformation causes a great challenge to current tracking algorithms. We propose a segmentation based approach to enable tracking can adapt to the varying shape of bird. The approach works by segmenting object at a region of interest, where is determined by the object localization method and heuristic edge information. The segmentation is performed by Markov random field, which is trained by foreground and background mixture Gaussian models. Experiments demonstrate that the proposed approach provides the ability to handle large deformations and outperforms the most state-of-the-art tracker in the infrared flying bird tracking problem.

A dual approach for machine-awareness in indoor environment combining pseudo-3D imaging and soft-computing techniques

3D spatial characterization of objects is a key-step for autonomous knowledge extraction leading to machine-awareness. In this paper we describe a dual-images-based approach for objects’ 3D characterization and localization for Machine-Awareness in indoor environment. The so-called dual images are provided by color and depth cameras of Kinect system, which presents an appealing potential for 3D objects modeling and localization. Placing the human–machine (including human-robot) interaction as a primary outcome of the intended visual Machine-Awareness in investigated system, we aspire proffering the machine spatial awareness about its surrounding environment with identification and semantically description of relative positions of detected items in that environment. In this research work, the aforementioned pseudo-3D imaging and Soft-computing techniques are combined in order to extract, to recognize and to spatially characterize objects and the distance between objects in 3D environment. In other words, on the one hand, a pseudo-3D object modeling and localization method based on the dual-images of Kinect system is proposed, in which the computation of object’s spatial characterization is presented. On the other hand, a computational semantically description algorithm by using Fuzzy Inference System is proposed, in which the numerical-to-semantic conversion for machine awareness is accomplished. Experimental results validating the investigated approach are reported and discussed.

Object Localization and Size Measurement Using Networked Address Event Representation Imagers

This letter presents a novel object localization and size measurement method using networked address event representation cameras. The algorithm uses circle of Apollonius to calculate results without computing trigonometric functions. Experiment results show that the localization and size error is under 5%. The algorithm has low communication and computation complexity; therefore, it fits low-power applications, such as identifying and tracking moving objects in the field.

Top-Down Visual Saliency via Joint CRF and Dictionary Learning

Top-down visual saliency is an important module of visual attention. In this work, we propose a novel top-down saliency model that jointly learns a Conditional Random Field (CRF) and a visual dictionary. The proposed model incorporates a layered structure from top to bottom: CRF, sparse coding and image patches. With sparse coding as an intermediate layer, CRF is learned in a feature-adaptive manner; meanwhile with CRF as the output layer, the dictionary is learned under structured supervision. For efficient and effective joint learning, we develop a max-margin approach via a stochastic gradient descent algorithm. Experimental results on the Graz-02 and PASCAL VOC datasets show that our model performs favorably against state-of-the-art top-down saliency methods for target object localization. In addition, the dictionary update significantly improves the performance of our model. We demonstrate the merits of the proposed top-down saliency model by applying it to prioritizing object proposals for detection and predicting human fixations.

BLSTM Recurrent Neural Network for Object Recognition

Multi-object relationship information can help eliminate some incorrect combinations or locations of objects. Moreover, it is favorable to extract scene information for object recognition. In this paper, we introduce a new way to generate image representation and propose a deep learning framework to fuse the contextual dependencies among objects and scene information in an image. It adopts a bidirectional long short-term memory recurrent neural network (BLSTM-RNN) to deal with the problem of variable-length sequence produced by local detectors in different images. Then it is applied to the existing tree context model for further recognition. Experimental results on SUN09 dataset show that our model outperforms the state-of the-art object localization methods.

Landmark constellation models for medical image content identification and localization.

Many medical imaging tasks require the detection and localization of anatomical landmarks, for example for the initialization of model-based segmentation or to detect anatomical regions present in an image. A large number of landmark and object localization methods have been described in the literature. The detection of single landmarks may be insufficient to achieve robust localization across a variety of imaging settings and subjects. Furthermore, methods like the generalized Hough transform yield the most likely location of an object, but not an indication whether or not the landmark was actually present in the image. For these reasons, we developed a simple and computationally efficient method combining localization results from multiple landmarks to achieve robust localization and to compute a localization confidence measure. For each anatomical region, we train a constellation model indicating the mean relative locations and location variability of a set of landmarks. This model is registered to the landmarks detected in a test image via point-based registration, using closed-form solutions. Three different outlier suppression schemes are compared, two using iterative re-weighting based on the residual landmark registration errors and the third being a variant of RANSAC. The mean weighted residual registration error serves as a confidence measure to distinguish true from false localization results. The method is optimized and evaluated on synthetic data, evaluating both the localization accuracy and the ability to classify good from bad registration results based on the residual registration error. Two application examples are presented: the identification of the imaged anatomical region in trauma CT scans and the initialization of model-based segmentation for C-arm CT scans with different target regions. The identification of the target region with the presented method was in 96% of the cases correct. The presented method is a simple solution for combining multiple landmark localization results. With appropriate parameters, outlier suppression clearly improves the localization performance over model registration without outlier suppression. The optimum choice of method and parameters depends on the expected level of noise and outliers in the application at hand, as well as on the focus on localization, classification, or both. The method allows detecting and localizing anatomical fields of view in medical images and is well suited to support a wide range of applications comprising image content identification, anatomical navigation and visualization, or initializing the pose of organ shape models.

Accurate Aerial Object Localization Using Gravity and Gravity Gradient Anomaly

Autonomous underwater vehicles (AUVs) have been widely used in diverse contexts, especially military affairs. The smooth operation of the AUV requires accurate localization of surrounding objects, especially the aerial objects. In this letter, a novel and practical method is presented for aerial object localization by using gravity and gravity gradient anomaly. Different from the state-of-the-art methods, such as GPS, radar, and laser, the proposed method runs in a passive manner and achieves AUV invisibility without energy emission. Compared with the object localization methods based on gravity and gravity gradient inversion, the proposed method is more practical as no large area gravity and gravity gradient measurements are needed to estimate the object mass. Experimental results demonstrate that the proposed method performs better than the existing methods.

An Event-Driven Object Localization Method Assisted by Beacon Mobility and Directional Antennas

An event-driven object localization method based on directional antennas is proposed in this paper. First the event occurrence spot is divided into four sections. Then low altitude UAV (Unmanned Aerial Vehicle) is employed to deploy DAWSN (Wireless Sensor Networks for Disaster Assistance) for urgent observation and communication. By means of a mobile anchor with four directional antennas and a GPS module, obstacle avoidance traverse in DAWSN is realized and the locations during mobility are broadcast. Unknown nodes take these locations as virtual anchors and project them onto a virtual motion path, and then the coordinates of unknown nodes are solved with extended directional localization method. This range-free method does not require plenty of anchor nodes and complicated computation. With small positioning error and large positionable node ratio (PNR), it allows the virtual anchor to move along any curve path and can be utilized under the event-driven scenario to provide self-localization for DAWSN.

An Object-Level High-Order Contextual Descriptor Based on Semantic, Spatial, and Scale Cues.

Context has been playing an increasingly important role in areas such as object detection, scene understanding, and image segmentation. Although many different types of contextual cues have been successfully explored, most of them only consider the pair-wise relationship between objects or parts. Several models utilize the high-order relationship for encoding contextual information. However, they mainly use a single contextual cue. In this paper, we present a novel high-order contextual descriptor (HOOD) to measure the strength of interactions among objects within an image. Heterogeneous contextual cues like semantic, spatial, and scale contexts are jointly integrated into HOOD to define the high-order interactions. The strength of these interactions are inferred by applying Bayes' rule on the pure dependence of the involved objects. As a result, an object-level graph is constructed to represent the contextually consistent interactions. Moreover, we propose a HOOD based object localization framework to verify the effectiveness of HOOD. Experimental results on two benchmark datasets including SUN09 and PASCAL2007 show that our framework outperforms the state-of-the-art context based object localization methods. Finally, we apply HOOD on two multimedia applications: structured image retrieval and out-of-context object detection, which demonstrates the potential usages of HOOD.

A Real-Time and Effective Object Recognition and Localization Method

In this paper, we realize object recognition and localization in a real time based on appearance features of object. For object recognition, we proposed to use global feauture (color) of images, and with an improved color image segmentation algorithm to realize threshold segmentation based on pixels in the image’s HSV color model by using the tool OpenCV, so we can realize the special color object recognition. Further the object can be localized with the ground constrained method by using the camera perspective geometry model. In the lab conditions, we realized single color object recognition and localization by transplanting the algorithm into Amigobots mobile robot and proved this method is simple, effective and real-time.

The object localization of semantic blocks on raster images

The method of object localization of semantic blocks on raster images is considered. The basis of an offered method are results of analysis of adaptive pyramidal representation of raster images. The feature of the algorithm of forming pyramidal representation is the relation between last level of hierarchy value with the level of semantic value that allows to find the areas representing the greatest interest in human perception and to use such selection for dynamic control of a compression.

Selective Search Method for Object Localization and Detection using Wavelets and Hierarchical Segmentations

This article proposes a selective search method for object localization in natural images by applying image multi-segmentation, image scaling, and heuristics. The method increases the number of generated windows that delimitate the area of an object with an accuracy superior to 50%. Over-segmentation is applied on original size images in order to locate small objects, and it is also applied over scaled images because these can still be over-segmented. This process produces less regions on areas with many textures. The over-segmentation was applied using the CIE Luv color model, and using the H and the I channels of the HSI model. The proposed method is category independent and allows the location of objects with heterogeneous characteristics by using heuristics and hierarchical segmentation. The proposed method produces 9, 366 windows per image covering 96.78% of the objects in the PASCAL VOC 2007 test image collection, increasing in 0.8% the localization results reported in the state of the art.

Localization in 2D Using Beacons of Low Frequency Magnetic Field

In this work we propose methods for object localization in 2D using beacons of low frequency quasi-static magnetic field. From a practical point of view, localization in 2D is sufficient for many applications, requiring much less calculations than in 3D, making it more robust and easier to implement in real-time low power applications. The low frequency magnetic field may penetrate foliage, soil, buildings, and many other types of media. This is an important advantage over traditional localization methods such as sonar or radar, where effective operation requires line-of-sight. Another advantage of the low frequency magnetic fields is that there is no direct influence by bad weather conditions and diurnal variations. Opposite to traditional electromagnetic methods, where operational range is usually more than a wavelength, low frequency induction approach results in a relatively limited localization range. Each beacon comprises a coil generating a magnetic field of a unique frequency in the ULF band. The generated magnetic fields are sensed by a search-coil magnetometer. The magnetometer readings are processed to estimate the magnitude and phase of the received beacons signals, which are used to localize the magnetometer. For a moving object, we propose to combine localization together with tracking algorithm using a data fusion approach. The proposed methods have been tested using numerous computer simulations, showing accurate localization results. A prototype was developed and used in field experiments, validating simulation results. The good accuracy together with a simple implementation makes the proposed methods attractive to many real-time low power field applications.

A method of 3D object recognition and localization in a cloud of points

The proposed method given in this article is prepared for analysis of data in the form of cloud of points directly from 3D measurements. It is designed for use in the end-user applications that can directly be integrated with 3D scanning software. The method utilizes locally calculated feature vectors (FVs) in point cloud data. Recognition is based on comparison of the analyzed scene with reference object library. A global descriptor in the form of a set of spatially distributed FVs is created for each reference model. During the detection process, correlation of subsets of reference FVs with FVs calculated in the scene is computed. Features utilized in the algorithm are based on parameters, which qualitatively estimate mean and Gaussian curvatures. Replacement of differentiation with averaging in the curvatures estimation makes the algorithm more resistant to discontinuities and poor quality of the input data. Utilization of the FV subsets allows to detect partially occluded and cluttered objects in the scene, while additional spatial information maintains false positive rate at a reasonably low level.

IMPLICIT SHAPE MODELS FOR OBJECT DETECTION IN 3D POINT CLOUDS

Abstract. We present a method for automatic object localization and recognition in 3D point clouds representing outdoor urban scenes. The method is based on the implicit shape models (ISM) framework, which recognizes objects by voting for their center locations. It requires only few training examples per class, which is an important property for practical use. We also introduce and evaluate an improved version of the spin image descriptor, more robust to point density variation and uncertainty in normal direction estimation. Our experiments reveal a significant impact of these modifications on the recognition performance. We compare our results against the state-of-the-art method and get significant improvement in both precision and recall on the Ohio dataset, consisting of combined aerial and terrestrial LiDAR scans of 150,000 m2 of urban area in total.

MULTIMODAL OBJECT LOCALIZATION

Various embodiments of the present invention are directed to systems and methods for multimodal object localization using one or more depth sensors and two or more microphones. In one aspect, a method comprises capturing three-dimensional images of a region of space wherein the object is located. The images comprise three-dimensional depth sensor observations. The method collects ambient audio generated by the object, providing acoustic observation regarding the ambient audio time difference of arrival at the audio sensors. The method determines a coordinate location of the object corresponding to the maximum of a joint probability distribution characterizing the probability of the acoustic observations emanating from each coordinate location in the region of space and the probability of each coordinate location in the region of space given depth sensor observations.

Keyframe-based recognition and localization during video-rate parallel tracking and mapping

Generating situational awareness by augmenting live imagery with collocated scene information has applications from game-playing to military command and control. We propose a method of object recognition, reconstruction, and localization using triangulation of SIFT features from keyframe camera poses in a 3D map. The map and keyframe poses themselves are recovered at video-rate by bundle adjustment of FAST image features in the parallel tracking and mapping algorithm. Detected objects are automatically labeled on the user's display using predefined annotations. Experimental results are given for laboratory scenes, and in more realistic applications.

A method for localizing seismic sources whose signals are separated by an order of difference from the background noise

A new method of endogenic object localization is suggested that applies the criterion of the degree of the relative order of complex oscillations based on over� normalized entropy. The method makes possible localization of deep sources, the signal from which can be distinguished over the background noise by the degree of chaos. The method does not need any infor� mation about the velocity model, while the measure� ments of the wave field are carried out sequentially, but not simultaneously, at the points of recording over the profile of the area, which is a principle difference of the new method from the traditional methods of source localization. The method is used for the local� ization of sources based on records of the seismic noise