Chamfer Matching Research Articles

TM3Loc: Tightly-Coupled Monocular Map Matching for High Precision Vehicle Localization

Vision-based map-matching with HD map for high precision vehicle localization has gained great attention for its low-cost and ease of deployment. However, its localization performance is still unsatisfactory in accuracy and robustness in numerous real applications due to the sparsity and noise of the perceived HD map landmarks. This article proposes the tightly-coupled monocular map-matching localization algorithm (TM3Loc) for monocular-based vehicle localization. TM3Loc introduces semantic chamfer matching (SCM) to model monocular map-matching problem and combines visual features with SCM in a tightly-coupled manner. By applying the sliding window-based optimization technique, the historical visual features and HD map constraints are also introduced, such that the vehicle poses are estimated with an abundance of visual features and multi-frame HD map landmark features, rather than with single-frame HD map observations in previous works. Experiments are conducted on large scale dataset of 15 km long in total. The results show that TM3Loc is able to achieve high precision localization performance using a low-cost monocular camera, largely exceeding the performance of the previous state-of-the-art methods, thereby promoting the development of autonomous driving.

Evaluation of the XVI dual registration tool for image-guided radiotherapy in prostate cancer.

To compare the reliability and the required time for two cone-beam CT (CBCT) registration methods for prostate irradiation (PI) and prostate bed irradiation (PBI). Two-hundred treatment fractions (in 10 PI and 10 PBI patients) were reanalyzed, using two CBCT registration methods: (1) a combination of an automated chamfer matching (CM) with manual matching (MM), and (2) the automated XVI dual registration tool (DRT). Bland-Altman 95% Limits of Agreement (LoA) were used to assess agreement with manual registration by Radiation Oncologists. All 95% LoA for CM+MM were≤0.33cm. For DRT, several 95% LoA were notably larger than the predefined clinical threshold of 0.3cm: -0.47 to +0.25cm (PI) and -0.36 to +0.23cm (PBI) for the superior-inferior direction and -0.52 to +0.24cm (PI) and -0.38 to +0.31cm (PBI) for the anterior-posterior direction.For PI, the average time required was 33s with CM+MM versus only 18s with DRT (p=0.002). For PBI, this was 13 versus 19s, respectively (p=0.16). For PI, DRT was significantly faster than CM+MM, but the accuracy is insufficient to use without manual verification. Therefore, manual verification is still warranted, but could offset the time benefit. For PBI, the CM+MM method was faster and more accurate.

Map Matching-Based Driving Lane Recognition for Low-Cost Precise Vehicle Positioning on Highways

This paper proposes a map matching-based driving lane recognition system for low-cost precise vehicle positioning on highways. The proposed method finds the position where the road boundaries detected by a LIDAR sensor are best matched with the road boundaries of the precise digital map and then recognizes the driving lane based on the position. To improve the limitations of the existing Chamfer matching method, this paper proposes the following two methods. The first is a method of generating one candidate position for each lane by using the lane information of the map and the left and right lane offsets obtained from the front camera. Second is a simple matching method that converts the LIDAR and map data into arrays and compares only the lateral distances to measure the matching score. In addition, this paper proposes a method for maintaining sufficient road boundary information in every frame by applying a temporal accumulation to the LIDAR data. In the experiment, the proposed method was quantitatively evaluated using a database acquired from sensors mounted on a real vehicle on highways. The experimental results show that the driving lane recognition rate is 99.56%, the lateral position error is 0.49m, and the processing time is 6.4ms. These results prove that the proposed system provides sufficient accuracy and reliability even in the presence of GPS position errors in various highway environments.

Accurate 6DOF Pose Tracking for Texture-Less Objects

A reliable and accurate visual object 6DoF pose tracking system for texture-less objects plays an important role in various fields of modern industry hence it has been a hot research topic for decades. Traditional feature-based pose tracking methods require rich feature points on objects, and it cannot handle texture-less objects. To tackle this problem, the paper proposes a novel edge-based method for continuous 6DOF pose tracking of texture-less objects. The pose of the object is estimated by minimizing the matching error between extracted image edges and re-projected CAD model edges. The matching error is represented using the Directional Chamfer Matching (DCM) Tensor. Compared with previous methods, the proposed system improves the overall performance of pose tracking system in two ways. Firstly, the method proposes an analytical mathematic model in the optimization process; Secondly, the method proposes an adaptive edge point weighting algorithm to tackle the occlusion or edge weakness problem. Both methods help improve the accuracy and robustness of the pose estimation system. With the benefit of GPU acceleration on DCM Tensor calculation the proposed method could run in real time. Extensive experiments are conducted both on public datasets and CG-rendered synthetic datasets to validate the accuracy and robustness of the proposed algorithm against other state-of-the-art methods.

3D Object Pose Estimation Using Chamfer Matching and Flexible CAD File Base

Estimating the object pose is an interesting topic in the industrial robotic vision field. By having an accurate result for detecting object pose, it means the system performs the task as the target in the bin-picking technique. The methods which are developed are varies widely. But the challenge for this paper is estimating a 3D object using mono camera accurately. The object which is used in this paper has the symmetric rotational shape, in this case is the sprayer. In this paper, the camera uses a tool from the Blender Software, such that the ground truth is measurable and it will be the reference for calculating the error. The applied algorithms of this paper are Border Line Extraction Algorithm utilized in the template generation step as the reference template, Directional Chamfer Matching for detecting the coarse pose, and Lavenberg-Marquardt Method to optimize the object pose result. The result achieves the average error of the coarse pose for x and y position (translation pose) are 2.05 mm and 0.71 mm. Meanwhile for the optimized pose, the average error for x and y position (translation pose) are 1.82 mm and 0.24 mm. Regarding the rotational pose, the average error of the rotational coarse pose with respect to x and z axis are 0.01 degree and 0.45 degree. Whereas the average error of the rotational optimized pose with respect to x and z axis are 2.88 degree and 0.82 degree.

Figure-ground segmentation based on class-independent shape priors

We propose a method to generate figure-ground segmentation by incorporating shape priors into the graph-cuts algorithm. Given an image, we first obtain a linear representation of an image and then apply directional chamfer matching to generate class-independent, nonparametric shape priors, which provide shape clues for the graph-cuts algorithm. We then enforce shape priors in a graph-cuts energy function to produce object segmentation. In contrast to previous segmentation methods, the proposed method shares shape knowledge for different semantic classes and does not require class-specific model training. Therefore, the approach obtains high-quality segmentation for objects. We experimentally validate that the proposed method outperforms previous approaches using the challenging PASCAL VOC 2010/2012 and Berkeley (BSD300) segmentation datasets.

A model for human activities recognition using partially occluded images

The demand for Human Activities Recognition (HAR) from still images and corresponding features categorization using appropriate classifier is never ending. HAR of partially occluded object using still image is highly challenging than extracting information from video due to the absence of any prior knowledge resembling frames stream .This new research domain dealing with the computer identification and subsequent classification of specific activities to develop understanding of human behaviors has diversified applied interests in surveillance cameras, security systems and automotive industry. We propose a new model for HAR with occluded missing part of human body in still image.(A template is built to complete the full appearance predictable by the system and their subsequent features classification is made.)The problems of HAR in still images are addressed using nonlinear Support Vector Machine classifier. The process of de-noising and chamfer matching are performed. This model is simulated with1200 still images of (64 x 128) pixels based on existing datasets such (INRIA and KTH). Using this model ,a recognition rate 86% is achieved for seven activities such as running, walking, jumping, clapping, jogging, boxing and waving .The excellent features of the results suggest that our method may constitute a basis for HAR with occlusion human body parts in still images promising for accurate features classification.

Global localization of 3D point clouds in building outline maps of urban outdoor environments

This paper presents a method to localize a robot in a global coordinate frame based on a sparse 2D map containing outlines of building and road network information and no location prior information. Its input is a single 3D laser scan of the surroundings of the robot. The approach extends the generic chamfer matching template matching technique from image processing by including visibility analysis in the cost function. Thus, the observed building planes are matched to the expected view of the corresponding map section instead of to the entire map, which makes a more accurate matching possible. Since this formulation operates on generic edge maps from visual sensors, the matching formulation can be expected to generalize to other input data, e.g., from monocular or stereo cameras. The method is evaluated on two large datasets collected in different real-world urban settings and compared to a baseline method from literature and to the standard chamfer matching approach, where it shows considerable performance benefits, as well as the feasibility of global localization based on sparse building outline data.

Improved chamfer matching method for surface mount component positioning

This study is concerned with the surface mount component positioning problem and an improved chamfer matching method based on iterative position and rotation angle estimation approach is proposed. Instead of performing matching of the image with different pre-specified angle templates pixel by pixel, as does in traditional chamfer matching methods, the gradient information of the distance transform image is incorporated into the chamfer matching method and reduced computational cost of the method is achieved. The iterative formulas to update the translation and rotation angle are derived by reference to the characteristic of rigid body motion. The criterions of search region establishment, seed point selection and terminal conditions are given. The effectiveness of the proposed method is verified by applying the method to component positioning with actual captured images.

Identifying designs from incomplete, fragmented cultural heritage objects by curve-pattern matching

The study of cultural heritage objects with embellished realistic and abstract designs made up of connected and intertwined curves crosscuts a number of related disciplines, including archaeology, art history, and heritage management. However, many objects, such as pottery sherds found in the archaeological record, are fragmentary, making the underlying complete designs unknowable at the scale of the sherd fragment. The challenge to reconstruct and study complete designs is stymied because (1) most fragmentary cultural heritage objects contain only a small portion of the underlying full design, (2) in the case of a stamping application, the same design may be applied multiple times with spatial overlap on one object, and (3) curve patterns detected on an object are usually incomplete and noisy. As a result, traditional curve-pattern matching algorithms, such as Chamfer matching, may perform poorly in identifying the underlying design. We develop a new partial-to-global curve matching algorithm to address these challenges and better identify the full design from a fragmented cultural heritage object. Specifically, we develop the algorithm to identify the designs of the carved wooden paddles of the Southeastern Woodlands from unearthed pottery sherds. A set of pottery sherds, curated at Georgia Southern University, are used to test the proposed algorithm, with promising results.

Model-based Pose Estimation for Texture-less Objects with Differential Evolution Algorithm

This paper proposes a novel object-tracking method to estimate three dimensions position of texture-less objects using one camera system and 3D model. The system uses efficient chamfer matching method to calculated distances between 2D edge templates of pose hypotheses with edges from the Canny edge query image. Differential Evolution algorithm uses those distances as inputs to ensure the close optimum results and find the most suitable position of objects. For initialization the exhaustive searching is employed. With the good initialization, a smaller searching space is set to guaranty the online tracking ability. The first results showed the potential of the method in solving object tracking and detection problem.

Local part chamfer matching for shape-based object detection

Chamfer matching is one of the elegant and powerful tools for shape-based detection in cluttered images. However, the chamfer matching methods, including oriented chamfer matching (OCM) and directional chamfer matching (DCM), tend to produce bad detections due to deformation of object shapes and cluttering in the scene. To improve detection accuracy of these chamfer matching methods, we propose local part oriented chamfer matching (LPOCM) and local part directional chamfer matching (LPDCM). First, shape templates and discriminative contour fragments are learned, and then a shape representation is built using a Markov random field (MRF). Finally, the template detection in an input image is formulated as an inference in the MRF. Experimental results for benchmark datasets including ETHZ Shape Classes, INRIA Horses and Weizmann Horses clearly demonstrate that the proposed LPOCM and LPDCM significantly improve the detection accuracy of OCM and DCM without sacrificing much time efficiency.

A probabilistic sparse skeleton based object detection

We present a Markov Random Field (MRF) based skeleton model for object shape and employ it in a probabilistic chamfer-matching framework for shape based object detection. Given an object category, shape hypotheses are generated from a set of sparse (coarse) skeletons guided by suitably defined unary and binary potentials at and between shape parts. The Markov framework assures that the generated samples properly reflect the observed or desired shape variability. As the model employs a sparsely sampled skeleton, the shape hypotheses are in the form of linear boundary segments; hence, matching can be performed using Directional Chamfer Matching. As the number of states that each MRF node can take is small, the matching process is efficient. Experiments with giraffe and swan categories of the ETHZ Dataset demonstrate that the method perform well in the case of articulated objects.

Vision-Based Excavator Detection and Tracking Using Hybrid Kinematic Shapes and Key Nodes

AbstractEnhancing workplace safety continues to be a major task in the construction industry. Approximately 75% of struck-by fatalities are caused by inappropriate spatial-temporal relationships between construction workers and heavy equipment. Construction safety can be improved if the location and movement of heavy equipment are tracked in real time. However, detecting and tracking heavy equipment with kinematic joints and changing poses, such as excavators, is still a challenge for vision-based sensing methods. This study proposes to detect and track excavators using stereo cameras based on hybrid kinematic shape and key node features. Specifically, templates of excavator components are synthesized for detection following kinematic constraints of each component. Thereafter, a fast directional chamfer matching algorithm is used to detect the excavator components, and the detected components are articulated at the key nodes. Finally, the three-dimensional positions of the key nodes are tracked through tr...

Sketch based Image Retrieval using Information Content of Orientation

Background/Objectives: This paper presents an image retrieval system using hand drawn sketches of images. Sketch is one of the convenient ways to represent the abstract shape of an object. The main objective is to perform retrieval of images using edge content by prioritizing the blocks based on information. Methods/Statistical Analysis: Entropy based Histogram of Gradients (HOG) method is proposed to prioritize the block. The method helps to pick the candidate blocks dynamically to compare with database images. Findings: The performance of the method has been evaluated using benchmark dataset of Sketch Based Image Retrieval (SBIR) with other methods like Indexable Oriented Chamfer Matching (IOCM), Context Aware Saliency (CAS-IOCM) and Histogram of Gradients (HOG). Comparing to these methods the number of relevant images retrieved is high for our approach. Application/Improvement: Knowledge based block selection method improves the performance of the existing method.

Leaf segmentation in plant phenotyping: a collation study

Image-based plant phenotyping is a growing application area of computer vision in agriculture. A key task is the segmentation of all individual leaves in images. Here we focus on the most common rosette model plants, Arabidopsis and young tobacco. Although leaves do share appearance and shape characteristics, the presence of occlusions and variability in leaf shape and pose, as well as imaging conditions, render this problem challenging. The aim of this paper is to compare several leaf segmentation solutions on a unique and first-of-its-kind dataset containing images from typical phenotyping experiments. In particular, we report and discuss methods and findings of a collection of submissions for the first Leaf Segmentation Challenge of the Computer Vision Problems in Plant Phenotyping workshop in 2014. Four methods are presented: three segment leaves by processing the distance transform in an unsupervised fashion, and the other via optimal template selection and Chamfer matching. Overall, we find that although separating plant from background can be accomplished with satisfactory accuracy ($$>$$>90 % Dice score), individual leaf segmentation and counting remain challenging when leaves overlap. Additionally, accuracy is lower for younger leaves. We find also that variability in datasets does affect outcomes. Our findings motivate further investigations and development of specialized algorithms for this particular application, and that challenges of this form are ideally suited for advancing the state of the art. Data are publicly available (online at http://www.plant-phenotyping.org/datasets) to support future challenges beyond segmentation within this application domain.

Automated 3D–2D registration of X-ray microcomputed tomography with histological sections for dental implants in bone using chamfer matching and simulated annealing

We propose a novel 3D–2D registration approach for micro-computed tomography (μCT) and histology (HI), constructed for dental implant biopsies, that finds the position and normal vector of the oblique slice from μCT that corresponds to HI. During image pre-processing, the implants and the bone tissue are segmented using a combination of thresholding, morphological filters and component labeling. After this, chamfer matching is employed to register the implant edges and fine registration of the bone tissues is achieved using simulated annealing.The method was tested on n=10 biopsies, obtained at 20 weeks after non-submerged healing in the canine mandible. The specimens were scanned with μCT 100 and processed for hard tissue sectioning. After registration, we assessed the agreement of bone to implant contact (BIC) using automated and manual measurements. Statistical analysis was conducted to test the agreement of the BIC measurements in the registered samples.Registration was successful for all specimens and agreement of the respective binary images was high (median: 0.90, 1.–3. Qu.: 0.89–0.91). Direct comparison of BIC yielded that automated (median 0.82, 1.–3. Qu.: 0.75–0.85) and manual (median 0.61, 1.–3. Qu.: 0.52–0.67) measures from μCT were significant positively correlated with HI (median 0.65, 1.–3. Qu.: 0.59–0.72) between μCT and HI groups (manual: R2=0.87, automated: R2=0.75, p<0.001).The results show that this method yields promising results and that μCT may become a valid alternative to assess osseointegration in three dimensions.

Local interfractional setup reproducibility for 2 individual head and neck supports in head and neck cancer patients

PurposeIn radiation therapy, head and neck (H&N) supports and thermoplastic masks are used to reproduce the setup of patients for H&N treatment. Individualized supports that include the shoulders may improve the immobilization of the upper thorax region. The purpose of this study was to compare the local misalignment of the supraclavicular region using a vacuum cushion H&N support to a more simple in-house modified, clinically standard H&N support. Methods and materialsTwo groups of 15 patients were evaluated: the first group of patients was positioned using a vacuum cushion as an individual head support and the second group with a modified Posifix headrest (MPH). A total of 316 cone beam computed tomography (CBCTs; ~ 10 scans per patient) were evaluated using a multiple region of interest registration protocol. Local setup errors were measured using chamfer matching on the CBCT scan to the planning CT scan for 9 bony structures (cervical vertebrae 1, 3, 5, and 7 [C7], lower jaw, hyoid bone, larynx, skull, and jugular notch). In this study, we compared the local residual misalignments of the bony structures and in particular those of the jugular notch and C7 as surrogates of the shoulders and thorax region. The workload was qualitatively evaluated on the basis of open interviews. ResultsThe significant differences in group mean, systematic error, and random error of the local residual misalignments between the 2 groups for jugular notch and C7 were equal or smaller than 0.5 mm and 0.1 degrees, and for the other 7 bony structures were equal to or smaller than 0.6 mm and 1.2 degrees (larynx). There were no large differences in workload. ConclusionsNo clinically relevant differences were found between a modified Posifix headrest and an individual vacuum cushion for H&N cancer patients in local posture change at the level of the clavicle and upper thorax.

Directional Chamfer Matching in 2.5 Dimensions

Directional chamfer matching (DCM) has shown good results in many areas such as object recognition and pose estimation. Currently DCM has been applied only for two-dimensional (2-D) matching. In this letter, we present a DCM scheme that utilizes depth in addition to 2-D input, which we refer to as 2.5D DCM. We show that in situations such as 3-D model-based pose estimation, depth information can be exploited to achieve robust performance. We apply the proposed method for human motion capture (HMC), using the Human Eva I dataset. We compare our approach with alternative methods used for HMC. Our results show that using depth information makes traditional DCM robust. Furthermore, the proposed method outperforms the alternatives used for HMC in state of the art systems.

A Joint-Constraint Model-Based System for Reconstructing Total Knee Motion

Comprehending knee motion is an essential requirement for studying the causes of knee disorders. In this paper, we propose a new 2-D-3-D registration system based on joint-constraint model for reconstructing total knee motion. The proposed model that contains bone geometries and an articulated joint mechanism is first constructed from multipostural magnetic resonance volumetric images. Then, the bone segments of the model are hierarchically registered to each frame of the given single-plane fluoroscopic video that records the knee activity. The bone posture is iteratively optimized using a modified chamfer matching algorithm to yield the simulated radiograph which is the best fit to the underlying fluoroscopic image. Unlike conventional registration methods computing posture parameters for each bone independently, the proposed femorotibial and patellofemoral joint models properly maintain the articulations between femur, tibia, and patella during the registration processes. As a result, we can obtain a sequence of registered knee postures showing smooth and reasonable physiologic patterns of motion. The proposed system also provides joint-space interpolation to densely generate intermediate postures for motion animation. The effectiveness of the proposed method was validated by computer simulation, animal cadaver, and in vivo knee testing. The mean target registration errors for femur, tibia, and patella were less than 1.5 mm. In particular, small out-of-plane registration errors [less than 1 mm (translation) and 2° (rotation)] were achieved in animal cadaver assessments.