The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
A vast majority of consumer cameras operate the rolling shutter mechanism, which often produces distorted images due to inter-row delay while capturing an image. Recent methods for monocular rolling shutter compensation utilize blur kernel, straightness of line segments, as well as angle and length preservation. However, they do not incorporate scene geometry explicitly for rolling shutter correction,...
Existing methods for 3D scene flow estimation often fail in the presence of large displacement or local ambiguities, e.g., at texture-less or reflective surfaces. However, these challenges are omnipresent in dynamic road scenes, which is the focus of this work. Our main contribution is to overcome these 3D motion estimation problems by exploiting recognition. In particular, we investigate the importance...
In this paper we propose a fast method for detecting the ground plane in 3D scenes for an arbitrary roll angle rotation of a stereo vision camera. The method is based on the analysis of the disparity map and its “V-disparity” representation. First, the roll angle of the camera is identified from the disparity map. Then, the image is rotated to a zero-roll angle position and the ground plane is detected...
Estimating human pose, shape, and motion from images and videos are fundamental challenges with many applications. Recent advances in 2D human pose estimation use large amounts of manually-labeled training data for learning convolutional neural networks (CNNs). Such data is time consuming to acquire and difficult to extend. Moreover, manual labeling of 3D pose, depth and motion is impractical. In...
This paper presents a new depth estimation method for multiview systems with arbitrary camera locations. The method exploits the graph cuts method, where vertices of the graph represent segments used for controlling the trade-off between the quality of depth maps and the time of estimation, while preserving the original resolution of a depth map. Moreover, the inter-view consistency of the depth maps,...
Traditional stereo matching approaches generally have problems in handling textureless regions, strong occlusions and reflective regions that do not satisfy a Lambertian surface assumption. In this paper, we propose to combine the predicted surface normal by deep learning to overcome these inherent difficulties in stereo matching. With the selected reliable disparities from stereo matching method...
In this paper we propose an efficient solution to jointly estimate the camera motion and a piecewise-rigid scene flow from an RGB-D sequence. The key idea is to perform a two-fold segmentation of the scene, dividing it into geometric clusters that are, in turn, classified as static or moving elements. Representing the dynamic scene as a set of rigid clusters drastically accelerates the motion estimation,...
This paper presents the temporal enhancement of the graph-based depth estimation method, designed for multiview systems with arbitrarily located cameras. The primary goal of the proposed enhancement is to increase the quality of estimated depth maps and simultaneously decrease the time of estimation. The method consists of two stages: the temporal enhancement of segmentation required in used depth...
Depth estimation and spatial awareness given a single monocular image is a challenging task for a computer as depth information is not retained when the 3D world is projected onto a 2D plane. Therefore, we must combine our prior knowledge with other monocular cues present in the image, such as occlusion, texture variations, and shadows to understand the depth of the image. In this paper, we present...
When a planar structure is observed from multiple views, the projections of its corresponding 3D points on each image are related by a homography. Its estimation is a key step in many computer vision tasks where either the rigid motion between views or a per-pixel image correspondence is sought. The vast majority of multi-view homography estimation techniques relies on matching a sparse set of point-to-point...
In order to fast register a camera into a 3D scene model under the Manhattan-World assumption, a method of matching corresponding 2D and 3D lines based on vanishing point is proposed in this paper. Firstly, this method detects line segments and estimates three orthogonal vanishing points to determine the local length of camera and the matrix from world to camera space. Afterwards, one line is drawn...
In this research we address the problem of depth estimation using a single motionless monocular camera. In our method we make no use of reference objects or marks in the image plane or on the ground apart from a one-off object used for horizon line detection; even this, however, is not necessary if a vanishing point detection algorithm is employed. Camera height is the only known parameter that is...
Tumor cellularity, the number of cells in the tumor, is an important tissue microstructural feature, which is useful for cancer diagnosis and cell number related treatment. Histopathological examination of tissues reveals the tissue microstructure hence permits to investigate cellularity, but is usually available only as a small sample or after resection. Diffusion-weighted MRI (DWI) is a non-invasive...
Estimation of tooth axis is needed for some clinical dental treatment. Existing methods require to segment the tooth volume from Computed Tomography (CT) images, and then estimate the axis from the tooth volume. However, they may fail during estimating molar axis due to that the tooth segmentation from CT images is challenging and current segmentation methods may get poor segmentation results especially...
The intensity of the light observed from every position and direction in a real scene can be modeled as a highdimensional field, namely the plenoptic function. This field codes the radiance information as a function of space, orientation, wavelength, and time. In the scope of depth estimation, several strategies have been developed to obtain a representation of the spatial structure of a scene. However,...
Nowadays lasers in the measurement industry are widely used. The accuracy of optical rangefinders, which detect a laser line directly from a taken frame, lies in precise laser line detection. Unfortunately, during the measurement process some laser line reflections can occur, laser colored objects are present on the scene or segmented laser line is too wide due to the laser saturation effect. These...
The presence of motion blur is unavoidable in hand-held cameras, especially in low-light conditions. In this paper, we address the inverse rendering problem of estimating the latent image, scene depth and camera motion from a set of differently blurred images of the scene. Our framework can account for depth variations, non-uniform motion blur as well as mis-alignments in the captured observations...
For super resolution (SR) of a 3D scene from multi-view low resolution (LR) images, subpixel registration is one of the main problems due to the difference of the depth for different objects, especially for hand-held uncalibrated cameras. In this paper we proposed a method to solve this problem by estimating the depth map for the LR images and treating the registration as a planar segmentation problem...
Conventional stereo matching or depth estimation algorithms always provide incomplete disparity map. These pixels without depth estimation in the map are named depth gaps. Weak texture and occluded areas are main source of depth gaps. We propose a novel method to assign good depth estimation on the areas above. Our algorithm combines state-of-art superpixel segmentation approach and linear filter...
Modern forest inventory is based on the accurate and precise characterization of the 3D structure of the forest. Although LiDAR (Light Detection and Ranging) is an effective tool to estimate forest parameters, when acquired from single view point it is not able to represent accurately the entire scene. To solve this problem, in this paper we present a method that integrates the terrestrial and airborne...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.