1,721,045 research outputs found
Recovering an indoor 3D layout with top-down semantic segmentation from a single image
No full textIn this paper, we propose a framework to recover a 3D cuboidal indoor scene with a novel detector-based semantic segmentation feature and a carefully-modified orientation map. We use those features to mimic the ability of humans to recognize a 3D layout from a single image. We define all the potentials in our model under a conditional random field formulation. Our experimental results show the effectiveness of our new features which complement the limitations of existing bottom-up geometric features while achieving the state-of-the-art layout accuracy on the indoor UIUC dataset. (C) 2015 Elsevier B.V. All rights reserved.1122sciescopu
GENERATING A STRATEGY FOR CONFIGURATION DETERMINATION : A MODULE OF VISION ALGORITHM COMPILER FOR OBJECT LOCALIZATION PROGRAM
No full text1
Design of coupled strong classifiers in AdaBoost framework and its application to pedestrian detection
No full textIn the AdaBoost framework, a strong classifier consists of weak classifiers connected sequentially. Usually the detection performance of the strong classifier can be improved increasing the number of weak classifiers used, but the improvement is asymptotic. To achieve further improvement we propose coupled strong classifiers (CSCs) which consist of multiple strong classifiers connected in parallel. Complementarity between the classifiers is considered for reducing intra- and inter-classifier correlations of exponential loss of weak classifiers in the training phase, and dynamic programming is used during the testing phase to compute efficiently the final object score for the coupled classifiers. In addition to CSC concept, we also propose using Aggregated Channel Comparison Features (ACCFs) that take the difference of feature values of Aggregated Channel Features (ACFs), enabling significant performance improvement. To show the effectiveness of our CSC concept, we apply our algorithm to pedestrian detection. Experiments are conducted using four well-known benchmark datasets based on ACFs, ACCFs, and Locally Decorrelated Channel Features (LDCFs). The experimental results show that our CSCs give better performance than the conventional single strong classifier for all cases of ACFs, ACCFs, and LDCFs. Especially our CSCs combined with ACCFs improve the detection performance significantly over ACE detector, and its performance is comparable to those of the state-of-the-art algorithms while using the simple ACE-based features. (C) 2015 Elsevier B.V. All rights reserved.11811sciescopu
Linear band detection based on the Euclidean distance transform and a new line segment extraction method
No full textA linear band, which is a straight line segment with some width (i.e., thickness), is a more structured, higher-level feature compared to edge or line features. In spite of the usefulness of linear bands as features, papers dealing with their detection problem are rare. In this paper, we propose a new method for detecting linear bands in gray-scale images. We first talk about our opinion on what types of linear bands a desirable detector should be able to detect, and then give a description on how we designed our detector to achieve the goal. Our method consists largely of two parts: (1) extracting the candidate center line pixels of the linear bands contained in an input gray-scale image (sub-parts: edge detection, Euclidean distance transform, ridge detection in a distance map, and noisy ridge pixel removal), (2) extracting line segments from the result of (1) using our new line segment detection method (sub-parts: modified Hough transform, base line segment grouping, redundant line segment removal, and postprocessing). Experimental results show that our method is practical and robust. (C) 2001 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.X1114sciescopu
Fast line segment grouping method for finding globally more favorable line segments
No full textIn this paper, we propose a new method for extracting line segments from edge images. Our method basically follows a line segment grouping approach. This approach has many advantages over a Hough transform based approach in practical situations. However, since the process of the conventional line segment grouping approach is purely local, it does not provide a mechanism for finding more favorable line segments from a global point of view. Our method overcomes the local nature of the conventional line segment grouping approach, while retaining most of its advantages, by incorporating the useful concept of the Hough transform based approach into the line segment grouping approach. Our method is fast and allows elementary line segments to be shared simultaneously by several line segments, and the degree of sharing is determined by a user-specified threshold. We performed a series of tests to compare the performance of our method with that of six other methods. Throughout the tests, our method ranked in the top two of the tested methods both in detection rate and computation time. (C) 2002 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.X1124sciescopu
Practical ways to calculate camera lens distortion for real-time camera calibration
No full textIn this paper, we address practical methods for calculating camera lens distortion for real-time applications. Although the lens distortion problem can be easily ignored for constant-parameter lenses, it becomes important in the field of real-time camera calibrations, particularly for zoom lenses. Tsai's camera calibration method, which is adopted in this paper for real-time application, consists of two stages. While some camera parameters can be calculated algebraically in the first stage, a nonlinear optimization process is involved in the second stage for calculating other parameters including lens distortion! which requires a large number of calculations. However, if the lens distortion can be calculated independently of the other camera parameters, we can easily calibrate a camera with a linear method without a computational burden. We propose two different methods for calculating lens distortion independently. These methods are so simple and require so few calculations that the lens distortion can be rapidly calculated even in real-time applications. The first one uses a look-up-table (LUT) of focal length and lens distortion, which can be constructed in the initialization process. The second one is a feature-based method rising the relationship between the feature points found in an image. Experiments were carried out for both methods, results of which show that the proposed methods are favorably comparable in performance with the non-real-time full optimization method. (C) 2001 Pattern Recognition Society. Published by Elsevier Science Ltd, All rights reserved.X1113sciescopu
A linear metric reconstruction by complex eigen-decomposition
No full textThis paper proposes a linear algorithm for metric reconstruction from projective reconstruction. Metric reconstruction problem is equivalent to estimating the projective transformation matrix that converts projective reconstruction to Euclidean reconstruct ion. We build a quadratic form froin dual absolute conic projection equation with respect to the elements of the transformation matrix. The matrix of quadratic form of rank 2 is then eigen-decomposed to produce a linear estimate. The algorithm is applied to three different sets of real data and the results show a feasibility of the algorithm. Additionally, our comparison of results of the linear algorithm to results of bundle adjustment, applied to sets of synthetic image data having Gaussian image noise, shows reasonable error ranges.open111sciescopu
A new graph cut-based multiple active contour algorithm without initial contours and seed points
No full textThis paper presents a new graph cut-based multiple active contour algorithm to detect optimal boundaries and regions in images without initial contours and seed points. The task of multiple active contours is framed as a partitioning problem by assuming that image data are generated from a finite mixture model with unknown number of components. Then, the partitioning problem is solved within a divisive graph cut framework where multi-way minimum cuts for multiple contours are efficiently computed in a top-down way through a swap move of binary labels. A split move is integrated into the swap move within that framework to estimate the model parameters associated with regions without the use of initial contours and seed points. The number of regions is also estimated as a part of the algorithm. Experimental results of boundary and region detection of natural images are presented and analyzed with precision and recall measures to demonstrate the effectiveness of the proposed algorithm.X1110sciescopu
Affine motion based CMOS distortion analysis and CMOS digital image stabilization
No full textThe CMOS image distortion is due to the rolling shutter in CMOS image sensors (CISs) and it can be more exaggerated when a CIS camera moves rapidly. Several methods have been proposed to remove CMOS distortions made by the translational motion. But, in this paper, we propose the affine motion based CMOS distortion correction method combined with digital image stabilization. To remove CMOS distortions due to the affine global image motion, of the CMOS distortion model is proposed to explain the effect affine global image motion on the CMOS distortion in CISs. To improve CMOS video's visuals, we propose CMOS digital image stabilization to remove the jittering motions in the new image sequence obtained by our correction method. In addition, to reduce the computational time and the outlier effect, a reliable feature selection method is proposed to be used in the affine global image motion estimation. In the experiment results, we show that the proposed CMOS distortion correction method is more general than previous ones. Also, we show that our stabilization method can improve CMOS video's visuals running in real-time(1).X1122sciescopu
Practical background estimation for mosaic blending with patch-based Markov random fields
No full textIn this paper, we present a new background estimation algorithm which effectively represents both background and foreground. The problem is formulated with a labeling problem over a patch-based Markov random field (MRF) and solved with a graph-cuts algorithm. Our method is applied to the problem of mosaic blending considering the moving objects and exposure variations of rotating and zooming camera. Also, to reduce seams in the estimated boundaries, we propose a simple exposure correction algorithm using intensities near the estimated boundaries. (c) 2008 Published by Elsevier Ltd.X117sciescopu
- …
