A network that learns to recognize three-dimensional objects

The visual recognition of three-dimensional (3-D) objects on the basis of their shape poses at least two difficult problems. First, there is the problem of variable illumination, which can be addressed by working with relatively stable features such as intensity edges rather than the raw intensity images. Second, there is the problem of the initially unknown pose of the object relative to the viewer. In one approach to this problem, a hypothesis is first made about the viewpoint, then the appearance of a model object from such a viewpoint is computed and compared with the actual image. Such recognition schemes generally employ 3-D models of objects, but the automatic learning of 3-D models is itself a difficult problem. To address this problem in computational vision, we have developed a scheme, based on the theory of approximation of multivariate functions, that learns from a small set of perspective views a function mapping any viewpoint to a standard view. A network equivalent to this scheme will thus 'recognize' the object on which it was trained from any viewpoint.     

Hearing visual motion in depth

Auditory spatial perception is strongly affected by visual cues. For example, if auditory and visual stimuli are presented synchronously but from different positions, the auditory event is mislocated towards the locus of the visual stimulus-the ventriloquism effect. This 'visual capture' also occurs in motion perception in which a static auditory stimulus appears to move with the visual moving object. We investigated how the human perceptual system coordinates complementary inputs from auditory and visual senses. Here we show that an auditory aftereffect occurs from adaptation to visual motion in depth. After a few minutes of viewing a square moving in depth, a steady sound was perceived as changing loudness in the opposite direction. Adaptation to a combination of auditory and visual stimuli changing in a compatible direction increased the aftereffect and the effect of visual adaptation almost disappeared when the directions were opposite. On the other hand, listening to a sound changing in intensity did not affect the visual changing-size aftereffect. The results provide psychophysical evidence that, for processing of motion in depth, the auditory system responds to both auditory changing intensity and visual motion in depth.     

图像序列中运动目标深度估计

通过对生物视觉深度运动知觉机理的研究，推导出基于深度运动知觉提取三维运动目标深度信息的数学模型和计算公式．研究结果表明，根据三维运动目标在二维成像平面上的运动可以判断是否存在深度方向变化以及大致区分目标运动的方向．同时如果假定已知运动目标的大小，根据它们在像平面上ｘ或ｙ方向的相对运动变化可以对于目标的深度运动进行估计．经过采用实际图像序列所进行的实验，获得了令人满意的结果．     

Efficient View-Based 3-D Object Retrieval via Hypergraph Learning

View-based 3-D object retrieval has become an emerging topic in recent years,especially with the fast development of visual content acquisition devices,such as mobile phones with cameras.Extensive research efforts have been dedicated to this task,while it is still difficult to measure the relevance between two objects with multiple views.In recent years,learning-based methods have been investigated in view-based 3-D object retrieval,such as graph-based learning.It is noted that the graph-based methods suffer from the high computational cost from the graph construction and the corresponding learning process.In this paper,we introduce a general framework to accelerate the learning-based view-based 3-D object matching in large scale data.Given a query object Q and one object O from a 3-D dataset D,the first step is to extract a small set of candidate relevant 3-D objects for object O.Then multiple hypergraphs can be constructed based on this small set of 3-D objects and the learning on the fused hypergraph is conducted to generate the relevance between Q and O,which can be further used in the retrieval procedure.Experiments demonstrate the effectiveness of the proposed framework.     

Interaction between colour and motion in human vision

There is a wealth of anatomical and psychological evidence which suggests that when people look at an object in the visual world, its various attributes such as colour, 'form', motion and depth are analysed by separate channels in the visual system. If so, how are these attributes put back together again to create a unified picture of the object? And if the object moves rapidly, how is perfect perceptual synchrony maintained between different features on its surface, if it is indeed true that they are being processed separately? Our evidence suggests that the visual system extracts certain conspicuous image features based on luminance contrast, and that the signals derived from these are then attributed to other features on the object, a process that we call 'capture'. Specifically, we find that when either illusory contours or random-dot patterns are moved in the vicinity of a colour-border, the colour border will also seem to move in the same direction even though it is physically stationary.     

Perception of shape from shading

The human visual system can rapidly and accurately derive the three-dimensional orientation of surfaces by using variations in image intensity alone. This ability to perceive shape from shading is one of the most important yet poorly understood aspects of human vision. Here we present several findings which may help reveal computational mechanisms underlying this ability. First, we find that perception of shape from shading is a global operation which assumes that there is only one light source illuminating the entire visual image. This implies that if two identical objects are viewed simultaneously and illuminated from different angles, then we would be able to perceive three-dimensional shape accurately in only one of them at a time. Second, three-dimensional shapes that are defined exclusively by shading can provide tokens for the perception of apparent motion, suggesting that the motion mechanism is remarkably versatile in the kinds of inputs it can use. Lastly, the occluding edges which delineate an object from its background can also powerfully influence the perception of three-dimensional shape from shading.     

计算机模拟NC加工的基础研究——空间实体“扫”形状的几何造型

为了利用计算机逼真地模拟NC加工,必须解决空间实体“扫”形状的几何造型问题。本文就求解空间实体的“扫”形状提出三种方法:①瞬间法;②垂直截断面法及③平行截断面法。并用这些方法构成的程序模块与TIPS-1系统相装配,调试结果,证实了所提出的方法是行之有效的。     

改进的动态过程三维面形重建技术

提出一种采用标记条纹进行跟踪的动态过程三维面形重建方法,能有效地解决因成像设备拍摄速度跟不上物体运动速度以及物体表面破裂等问题对动态过程三维面形重建所带来的影响,从而获得物体运动过程中正确的三维面形分布．计算机模拟和实验证实了该种方法的正确性．     

Self-motion and the perception of stationary objects

One of the ways that we perceive shape is through seeing motion. Visual motion may be actively generated (for example, in locomotion), or passively observed. In the study of the perception of three-dimensional structure from motion, the non-moving, passive observer in an environment of moving rigid objects has been used as a substitute for an active observer moving in an environment of stationary objects; this 'rigidity hypothesis' has played a central role in computational and experimental studies of structure from motion. Here we show that this is not an adequate substitution because active and passive observers can perceive three-dimensional structure differently, despite experiencing the same visual stimulus: active observers' perception of three-dimensional structure depends on extraretinal information about their own movements. The visual system thus treats objects that are stationary (in an allocentric, earth-fixed reference frame) differently from objects that are merely rigid. These results show that action makes an important contribution to depth perception, and argue for a revision of the rigidity hypothesis to incorporate the special case of stationary objects.     

用单个摄像机获取三维视觉信息的一种方法

讨论了在平面坐标系上摄像机的标定和利用单个摄像机获取三维视觉信息的一种方法。实验结果表明，采用单个摄像机获取三维视觉信息的精度与摄像机相对目标物体的高度及对象物本身的厚度有关，在视场范围内，对象物的厚度越小，摄像机坐标中心相对于目标物体表面中心的距离越远，定位精度越好。     

一种新的物体连续切片图象的截面三维重建

描述了一种新的截面三维重建方法,它是在截面重建法的基础上,引入仿射变换作为旋转、投影变换,利用线性插值消隐填充算法,生成物体的可视侧面,直接利用重建图象的深度信息计算可视侧面的灰度,在显示平面获得物体可视侧面的三维结构及形状信息.并可通过旋转投影变换获得物体任一可视侧面的重建显示图象.     

Bottlenose dolphins perceive object features through echolocation

How organisms (including people) recognize distant objects is a fundamental question. The correspondence between object characteristics (distal stimuli), like visual shape, and sensory characteristics (proximal stimuli), like retinal projection, is ambiguous. The view that sensory systems are 'designed' to 'pick up' ecologically useful information is vague about how such mechanisms might work. In echolocating dolphins, which are studied as models for object recognition sonar systems, the correspondence between echo characteristics and object characteristics is less clear. Many cognitive scientists assume that object characteristics are extracted from proximal stimuli, but evidence for this remains ambiguous. For example, a dolphin may store 'sound templates' in its brain and identify whole objects by listening for a particular sound. Alternatively, a dolphin's brain may contain algorithms, derived through natural endowments or experience or both, which allow it to identify object characteristics based on sounds. The standard method used to address this question in many species is indirect and has led to equivocal results with dolphins. Here we outline an appropriate method and test it to show that dolphins extract object characteristics directly from echoes.     

基于仿射不变性特征的视点空间划分

为了简化三维物体的识别过程,提高三维物体识别的识别率,该文利用Multi-scale autoconvolution、Trace变换、Zernike矩3种仿射不变性特征,对飞机、汽车、人等三维物体进行视点空间划分,用尽可能少的不等间隔的三维物体的二维投影图像来表达三维物体,并以此为依据进行三维物体识别。在此基础上提出一种针对不同类型物体的仿射不变性特征提取策略,并建立一个实现三维物体任意姿态识别的软件系统平台,应用Princeton形状标准库中的部分模型对该平台进行测试。结果表明,该方法能够取得较好的识别效果,识别率在90%以上。     

微孔结构对多孔材料细观力学特性的影响

在三维孔隙材料如泡沫材料中，胞孔的形状及大小并非完全均匀，因此借助二维均匀化方法来讨论胞孔形状及大小对多孔材料性能的影响。在线弹性范围内，根据均匀化理论，基于虚位移原理，结合有限元方法推导出二维周期性结构的均匀化的有限元格式。取具有不同孔洞形状的正方形胞元作为周期性结构的代表胞元，将三维孔隙材料简化为截面上具有规则孔洞的二维结构，来计算不同微孔形状及大小下的等效弹性常数；比较分析了微孔结构对多孔材料等效弹性常数的影响。计算结果分析表明，多孔材料的等效弹性参数不仅取决于微孔结构的数量，而且对微孔结构的形状也有一定程度的敏感，同时对基体材料的泊松比的变化敏感与否也与微孔结构有关。     

The role of disparity-sensitive cortical neurons in signalling the direction of self-motion

Movement of an observer through the environment generates motion on the retina. This optic flow provides information about the direction of self-motion, but only if it contains differential motion of elements at different depths. If the observer tracks a stationary object while moving in a direction different from his line of sight, the images of objects in the foreground and in the background move in opposite directions. We have found neurons in the cerebral cortex of monkeys that prefer one direction of motion when the disparity of a stimulus corresponds to foreground motion and prefer the opposite direction when the disparity corresponds to background motion. We propose that these neurons contribute a signal about the direction of self-motion.     

Does the brain know the physics of specular reflection?

Images of artificial and natural scenes typically contain many highlights generated by mirror-like reflection from glossy surfaces. Until recently, computational models of visual processes have tended to regard highlights as obscuring the structure of the underlying scene. The truth is that, on the contrary, highlights are rich in local geometric information. Here we report that the three-dimensional appearance of a highlight on a computer-simulated stereoscopic curved surface affects observers' judgment of surface gloss. We also show that the 3-D appearance of a highlight affects the perception of surface curvature--that is, it can force an ambiguous convex-concave figure to change state. We thus conclude that human visual analysis seems to employ a physical model of the interaction of light with curved surfaces, a model firmly based on ray optics and differential geometry.     

Detection of salient objects with focused attention based on spatial and temporal coherence

The understanding and analysis of video content are fundamentally important for numerous applications,including video summarization,retrieval,navigation,and editing.An important part of this process is to detect salient (which usually means important and interesting) objects in video segments.Unlike existing approaches,we propose a method that combines the saliency measurement with spatial and temporal coherence.The integration of spatial and temporal coherence is inspired by the focused attention in human vision.In the proposed method,the spatial coherence of low-level visual grouping cues (e.g.appearance and motion) helps per-frame object-background separation,while the temporal coherence of the object properties (e.g.shape and appearance) ensures consistent object localization over time,and thus the method is robust to unexpected environment changes and camera vibrations.Having developed an efficient optimization strategy based on coarse-to-fine multi-scale dynamic programming,we evaluate our method using a challenging dataset that is freely available together with this paper.We show the effectiveness and complementariness of the two types of coherence,and demonstrate that they can significantly improve the performance of salient object detection in videos.     

基于人类视觉系统的物体提取方案

提出了一种物体提取方案，根据人类视觉系统（ＨＶＳ）具有形状和运动两个并行通道的特点以及两者交互作用的原理，采用相对简单的图像分割和运动分割技术，将两者结合在一起，来增强对运动物体的分割处理     

航空螺旋桨振动特性影响因素

选取某航空活塞发动机的两桨叶定距螺旋桨为分析对象,基于通用有限元软件平台建立螺旋桨整体结构、单个桨叶结构的三维有限元模型。根据航空活塞动力装置性能参数,计算出螺旋桨在地面起飞状态下受到的气动、离心载荷,再利用有限元软件计算了螺旋桨整体结构、单个桨叶结构在无外载荷作用下的、只有气动载荷作用的、只有离心载荷作用的、气动和离心载荷同时作用的自振频率和振型。通过对计算结果的分析,发现了气动载荷、离心载荷、形状（整体或单个桨叶）因素对频率、振型的影响规律,对其他与气体有相互作用的旋转部件的振动特性计算和分析具有重要的指导意义。     

2-D shape completion with shape priors

We propose new techniques for 2-D shape/contour completion, which is one of the important research topics related to shape analysis and computer vision, e.g. the detection of incomplete objects due to occlusion and noises. The purpose of shape completion is to find the optimal curve segments that fill the missing contour parts, so as to acquire the best estimation of the original complete object shapes. Unlike the previous work using local smoothness or minimum curvature priors, we solve the problem under a Bayesian formulation taking advantage of global shape prior knowledge. With the priors, our methods are expert in recovering significant shape structures and dealing with large occlusion cases. There are two different priors adopted in this paper: (i) A generic prior model that prefers minimal global shape transformation (including non-rigid deformation and affine transformation with respect to a reference object shape) of the recovered complete shape; and (ii) a class-specific shape prior model learned from training examples of an object category, which prefers the reconstructed shape to follow the learned shape variation models of the category. Efficient contour completion algorithms are suggested corresponding to the two types of priors. Our experimental results demonstrate the advantage of the proposed shape completion approaches compared to the existing techniques, especially for objects with complex structure under severe occlusion.