Neural Mesh Fusion: Unsupervised 3D Planar Surface Understanding

Time Optimal Control for a Non-Linear Planar Vehicle Subject to Disturbances

Time optimal control of a non-linear surface vehicle subject to disturbances

Book published

The problem of an autonomous agent moving on a planar surface, such as an aerial drone or a small naval vessel can be treated as navigation between a series of points. In addition the robots operating in the real world often have to come up with a sequence of actions, a plan, which will take them from an initial state to the desired goal state. During planning, the robots have to take into account both discrete and continuous changes, as well as temporal constraints. While nominally the movement between each pair of points can be treated as a 1D projection of the movement on the vector connecting the two points, in the presence of disturbances, the full problem on the plane must be considered. The minimum-time optimal solution depends on the value and direction of the disturbance which in this paper is assumed to be a sum of a constant velocity of the medium (wind or current, respectively) and a bounded time-varying part with uncertain but bounded time derivative. We address the minimum time problem of a movement on a 2D plane with quadratic drag, under norm state (inertial vessel velocity), and norm control (acceleration) constraints. The worst case influence of an uncertain disturbance is evaluated by consideration of a differential game with state and control constraints. The structure and properties of the optimal solution were found and analyzed.

This paper deals with plane detection from a monocular image sequence without camera calibration or a priori knowledge about the egomotion. Within a framework of driver assistance applications, it is assumed that the 3D scene is a set of 3D planes. In this paper, the vision process considers obstacles, roads and buildings as planar structures. These planes are detected by exploiting iso-velocity curves after optical flow estimation. A Hough Transform-like frame called c-velocity was designed. This paper explains how this c-velocity, defined by analogy to the v-disparity in stereovision, can represent planes, regardless of their orientation and how this representation facilitates plane extraction. Under a translational camera motion, planar surfaces are transformed into specific parabolas of the c-velocity space. The error and robustness analysis of the proposed technique confirms that this cumulative approach is very efficient for making the detection more robust and coping with optical flow imprecision. Moreover, the results suggest that the concept could be generalized to the detection of other parameterized surfaces than planes.

As a classic project for the inheritance of Chinese folk culture, Wupaolong has the characteristics of unique material selection and complex production technology. However, as the finished products of folk handicrafts may not comply with production standards, a common problem in the production of handicrafts. Focusing on the problems of poor visual effects such as uneven surface and jagged edges in the craft paper binding model constructed by point cloud, and the need to enhance the sense of reality, this paper proposes an optimization method for 3D model and triangulation based on 3D line features. This method first preprocesses the feature points and performs sparse matching and dense matching on the processed feature points to generate a dense point cloud. The dense point cloud is then processed by an adaptive weighted median filtering method. This is followed by application of surface modeling method based on Delaunay algorithm to realize the reconstruction of object triangulation. Based on the triangular patch, the reference plane of the craft paper binding surface is fitted, and the model surface is corrected and optimized based on the reference plane. With the aid of three-dimensional line segments, the triangulation at the edge of the craft paper binding is corrected and optimized. Finally, an optimized three-dimensional model of craft paper binding is obtained. Experimental results show that the proposed algorithm can improve the unevenness of the 3D model plane and edges, and maintain the characteristics of the 3D model plane and edges. Compared with the existing surface reconstruction algorithms based on point cloud features, the optimized triangulation and model have higher accuracy and quality, which can greatly improve the visual effect of the model and improve the accuracy of the three-dimensional model. The research results of this paper can realize 3D modeling development for the production process of Wupaolong, effectively improve the elevation accuracy of the model, realize the production standard of digital Wupaolong model, and provide reference for folk production process.

Local dose analysis to predict acute and late urinary toxicities after prostate cancer radiotherapy: Assessment of cohort and method effects

Classical volume rendering is computed by casting a bundle of parallel rays from a flat viewing plane onto the volume data set, and produces as such a spatially limited view of the objects in the data set. The method described in this paper is able to generate an overall planar view of an object that is topologically compatible with the sphere, by firing rays from a nearby surrounding surface and by unfolding this surface in a 2D plane, without introducing major distortions. It has been devised to facilitate the interpretation of the cerebral cortex. An initial surface consisting of two hemi-ellipsoids, one to cover the top and another one to surround the bottom of the brain, is interactively defined and deformed via a deformable model approach towards a dilated version of the cortical surface of the brain. During deformation, the nodes on the surface are continuously redistributed, to maintain a near homothetic mapping with the plane. Once the surface has converged to the dilated brain surface, rays are casted from the nodes, according to the normal of the surface at the node. The shading result, computed at the intersection of the rays with the original brain surface, is mapped via the near homothetic mapping to the plane. With this approach sulci can be followed in their entirety, so that it is much easier to derive their spatial relationship and to recognize them.

This paper describes a surface flattening technique, which has been developed in particular to obtain a complete view of the cortical surface of the brain. However, the method is able to produce an overall planar view of any anatomical or real-life object, provided it is topologically compatible with the sphere (i.e. genus 0). It computes the shading of the original surface for rays casted from a nearby surrounding surface and unfolds this surface in a 2D plane, without introducing major distortions. The flat image consisting of the mapped shading results has the advantage that the sulci (i.e. the grooves characterizing the superficial brain geometry) of the cortical surface of the brain can be followed in their entirety, which facilitates the study and the recognition of their patterns. The new visualization method is integrated into a versatile medical image analysis environment. A first study to assess its usefulness has been accomplished and is also reported in this paper.

Computer animation of outdoor scenes is enhanced by realistic clouds. I discuss several different modelling and rendering schemes for clouds, and show how they evolved in my animation work. These include transparency-textured clouds on a 2D plane, smooth shaded or textured 3D cloud surfaces, and 3D volume rendering. For the volume rendering, I present various illumination schemes, including the density emitter, single scattering, and multiple scattering models. >

This study aimed to develop functional underwear for elderly women in their sixties in terms of good fit, wear comfort and body temperature regulation. To satisfy elderly women's physical and metabolical needs, an automatic temperature control system via PCM treatment was applied. Underwear pattern was produced by producing body surface replica, which was derived from 3D body parametric model. Differential ratios of outline length and area between 3D surface and 2D plane were 1.4% and 0.5%, respectively. The reduction rate was determined as 10% through the expert's evaluation. PCM treated fabric showed higher Q-max, meaning that it can facilitate the thermal transition in hot situation. Moreover, it also showed higher insulation to preserve heat and keep warm microclimate in a cold weather. Heat distribution measurements on various body parts revealed that the temperature after PCM treatment was significantly higher. The clothing pressure after 10% pattern reduction showed higher before reduction, at the same time, even lower than the comfort clothing pressure range of <TEX>$5{\sim}10gf/cm^2$</TEX>, implying that experimental garment of this research is acceptable in terms of clothing pressure. Evaluation results on the comfort to move in various motions proved that adequate clothing pressure improved the wear comfort in various motions.

Food volume estimation is an essential step in the pipeline of dietary assessment and demands the precise depth estimation of the food surface and table plane. Existing methods based on computer vision require either multi-image input or additional depth maps, reducing convenience of implementation and practical significance. Despite the recent advances in unsupervised depth estimation from a single image, the achieved performance in the case of large texture-less areas needs to be improved. In this paper, we propose a network architecture that jointly performs geometric understanding (i.e., depth prediction and 3D plane estimation) and semantic prediction on a single food image, enabling a robust and accurate food volume estimation regardless of the texture characteristics of the target plane. For the training of the network, only monocular videos with semantic ground truth are required, while the depth map and 3D plane ground truth are no longer needed. Experimental results on two separate food image databases demonstrate that our method performs robustly on texture-less scenarios and is superior to unsupervised networks and structure from motion based approaches, while it achieves comparable performance to fully-supervised methods.

Food volume estimation is an essential step in the pipeline of dietary assessment and demands the precise depth estimation of the food surface and table plane. Existing methods based on computer vision require either multi-image input or additional depth maps, reducing convenience of implementation and practical significance. Despite the recent advances in un- supervised depth estimation from a single image, the achieved performance in the case of large texture-less areas needs to be improved. In this paper, we propose a network architecture that jointly performs geometric understanding (i.e., depth prediction and 3D plane estimation) and semantic prediction on a single food image, enabling a robust and accurate food volume estimation regardless of the texture characteristics of the target plane. For the training of the network, only monocular videos with semantic ground truth are required, while the depth map and 3D plane ground truth are no longer needed. Experimental results on two separate food image databases demonstrate that our method performs robustly on texture-less scenarios and is superior to unsupervised networks and structure from motion based approaches, while it achieves comparable performance to fully-supervised methods.

In this paper, an intra prediction method is proposed for coding of depth pictures using plane modelling. Each pixel in the depth picture is related to the distance from a camera to an object surface, and pixels corresponding to a flat surface of an object form a relationship with the 2D plane surface. The plane surface can be represented by a simple equation in the 3D camera coordinate system in such a way that the coordinate system of depth pixels can be transformed to the camera coordinate system. This paper finds the parameters which define the plane surface closest to given depth pixels. The plane model is then used to predict the depth pixels on the plane surface. A depth prediction method is also devised for efficient intra prediction of depth pictures, using variable-size blocks. For prediction with variable-size blocks, the plane surface that occupies a large part of the picture can be predicted using a large block size. The simulation results of the proposed method show that the mean squared error is reduced by up to 96.6% for a block size of 4 × 4 pixels and reduced by up to 98% for a block size of 16 × 16, compared with the intra prediction modes of H.264/AVC and H.265/HEVC.

Abstract. In this paper, we proposed a new refinement procedure for the semi-global dense image matching. In order to remove outliers and improve the disparity image derived from the semi-global algorithm, both the local smoothness constraint and point cloud segments are utilized. Compared with current refinement technique, which usually assumes the correspondences between planar surfaces and 2D image segments, our proposed approach can effectively deal with object with both planar and curved surfaces. Meanwhile, since 3D point clouds contain more precise geometric information regarding to the reconstructed objects, the planar surfaces identified in our approach can be more accurate. In order to illustrate the feasibility of our approach, several experimental tests are conducted on both Middlebury test and real UAV-image datasets. The results demonstrate that our approach has a good performance on improving the quality of the derived dense image-based point cloud.