Rank estimation in 3D multibody motion segmentation

A novel technique for rank estimation in 3D multibody motion segmentation is proposed. It is based on the study of the frequency spectra of moving rigid objects and does not use or assume a prior knowledge of the objects contained in the scene (i.e. number of objects and motion). The significance of rank estimation on multibody motion segmentation results is shown by using two motion segmentation algorithms over both synthetic and real data.     

3D structure extraction coding of image sequences

This paper presents a 3D structure extraction coding scheme that first computes the 3D structural properties such as 3D shape, motion, and location of objects and then codes image sequences by utilizing such 3D information. The goal is to achieve efficient and flexible coding while still avoiding the visual distortions through the use of 3D scene characteristics inherent in image sequences. To accomplish this, we present two multiframe algorithms for the robust estimation of such 3D structural properties, one from motion and one from stereo. The approach taken in these algorithms is to successively estimate 3D information from a longer sequence for a significant reduction in error. Three variations of 3D structure extraction coding are then presented — 3D motion interpolative coding, 3D motion compensation coding, and “viewpoint” compensation stereo image coding — to suggest that the approach can be viable for high-quality visual communications.     

基于人工智能的足迹识别与特征提取

针对战场感知及侦破现场中传统人工主观经验检验与识别模式误差较大的问题,提出了一种基于人工智能的足迹识别与特征提取方法。采用三维形貌重构系统进行足迹图像采集,并将数字图像处理算法与传统足迹检验法结合,提取足迹的区域关系特征和形状长度特征,进而采用支持向量机的模式识别方法对提取的特征进行立体足迹身份鉴别对比实验。实验结果表明,所提方法准确率超过人工鉴别准确率,达到99.1%,可应用于战场感知及侦破现场足迹准确检测与识别,也可推广应用于人体身份鉴别的相关领域。     

Adaptive correction of nonuniform illumination of multispectral digital images

The majority of known methods for correction of multispectral images distorted by nonuniform illumination use the following distortion model: a certain part of the scene close to a directed light source is illuminated much brighter than the rest of the scene. However, another serious problem often arises in practice in the case of nonuniform illumination of 3D objects of the scene: extended heavily shadowed regions with a small area of transition from light to shadow are formed. In this study, a method for the locally adaptive correction of nonuniform illumination of multispectral digital images, which is based on an algorithm that imitates human visual perception, is proposed. The performance of the proposed method is compared to that of available algorithms for correction of color images distorted by both nonuniform illumination and the presence of shadow regions.     

Rigid and non-rigid 3D motion estimation from multiview image sequences

Multiview image sequence processing has been the focus of considerable attention in recent literature. This paper presents an efficient technique for object-based rigid and non-rigid 3D motion estimation, applicable to problems occurring in multiview image sequence coding applications. More specifically, a neural network is formed for the estimation of the rigid 3D motion of each object in the scene, using initially estimated 2D motion vectors corresponding to each camera view. Non-linear error minimization techniques are adopted for neural network weight update. Furthermore, a novel technique is also proposed for the estimation of the local non-rigid deformations, based on the multiview camera geometry. Experimental results using both stereoscopic and trinocular camera setups illustrate and evaluate the proposed scheme.     

Transform-based image enhancement algorithms with performancemeasure

This paper presents a new class of the "frequency domain"-based signal/image enhancement algorithms including magnitude reduction, log-magnitude reduction, iterative magnitude and a log-reduction zonal magnitude technique. These algorithms are described and applied for detection and visualization of objects within an image. The new technique is based on the so-called sequency ordered orthogonal transforms, which include the well-known Fourier, Hartley, cosine, and Hadamard transforms, as well as new enhancement parametric operators. A wide range of image characteristics can be obtained from a single transform, by varying the parameters of the operators. We also introduce a quantifying method to measure signal/image enhancement called EME. This helps choose the best parameters and transform for each enhancement. A number of experimental results are presented to illustrate the performance of the proposed algorithms.     

基于运动区域检测的会议视频清晰度增强处理

视频会议场景对视频增强的实时性有较高的要求。针对现有视频增强算法(如BM3D等)存在的耗时长的问题,基于变化区域检测提出一种基于运动区域检测的视频增强技术。对时序帧数据进行颜色空间转换,快速地将视频场景分为静止区域和运动区域,之后对静止区域进行时域降噪和视频增强。在视频帧集合的实验结果表明,该算法可以显著地增强图像纹理细节,并具有很强的时效性,在视频会议场景有很大的应用潜力。     

3D reconstruction from stereo images for interactions between real and virtual objects

The 3D reconstruction algorithm in a stereo image pair for realizing mutual occlusion and interactions between the real and virtual world in an image synthesis is proposed. A two-stage algorithm, consisting of disparity estimation and regularization is used to locate a smooth and precise disparity vector. The hierarchical disparity estimation technique increases the efficiency and reliability of the estimation process, and edge-preserving disparity field regularization produces smooth disparity fields while preserving discontinuities that result from object boundaries. Depth information concerning the real scene is then recovered from the estimated disparity fields by stereo camera geometry. Simulation results show that the proposed algorithm provides accurate and spatially correlated disparity vector fields in various types of images, and the reconstructed 3D model produces a natural space in which the real world and virtual objects interact with each other as if they were in the same world.     

人工光源条件下夜间雾天图像建模及去雾

夜间有雾图像光照不均匀,整体亮度较低,色偏严重,且人工光源周围存在光晕。现有的去雾模型和算法大多针对白天图像,其并不适用于夜间场景,夜间图像去雾颇具挑战性。该文深入分析夜间有雾图像的成像规律,建立含有人工光源的夜间雾天图像成像新模型,并在此基础上提出夜间图像去雾新算法。针对夜间图像光照不均问题,提出基于低通滤波的环境光估计方法,利用估计出的环境光可准确预测夜间场景传输率;针对目前夜间图像去雾后存在光源光晕问题,提出根据图像色度估计场景点属于近光源区域的程度,使算法能自适应地处理光源区域和非光源区域;针对非一致色偏问题,利用直方图匹配方法进行颜色校正。对大量图像进行实验,并与现有白天、夜晚图像去雾算法进行比较,验证了该文提出的夜间雾天图像成像模型及去雾算法的有效性。     

Plane surface detection and reconstruction using segment-based tensor voting

A Segment-based Tensor Voting (SBTV) algorithm is presented for planar surface detection and reconstruction of man-made objects. Our work is inspired by piecewise planar stereo reconstruction. During the vital procedure to detect and label the planar surface, the two main contributions are: first, tensor voting is used for obtaining the geometry attribute of the 3D points cloud. The candidate planar patches are generated through scene image segment of low variation of color and intensity. Second, we over-segment the scene image into the segment and the candidate 3D planar patch is generated. The SBTV algorithm is used on 3D points cloud sets to identify the co-plane on the candidate patch. After detecting every planar patch, the geometry architecture of object is obtained. The experiments demonstrate the effectiveness of our proposed approach on either outdoor or indoor datasets.     

Localizing text in scene images by boundary clustering, stroke segmentation, and string fragment classification

In this paper, we propose a novel framework to extract text regions from scene images with complex backgrounds and multiple text appearances. This framework consists of three main steps: boundary clustering (BC), stroke segmentation, and string fragment classification. In BC, we propose a new bigram-color-uniformity-based method to model both text and attachment surface, and cluster edge pixels based on color pairs and spatial positions into boundary layers. Then, stroke segmentation is performed at each boundary layer by color assignment to extract character candidates. We propose two algorithms to combine the structural analysis of text stroke with color assignment and filter out background interferences. Further, we design a robust string fragment classification based on Gabor-based text features. The features are obtained from feature maps of gradient, stroke distribution, and stroke width. The proposed framework of text localization is evaluated on scene images, born-digital images, broadcast video images, and images of handheld objects captured by blind persons. Experimental results on respective datasets demonstrate that the framework outperforms state-of-the-art localization algorithms.     

基于空间金字塔匹配的单目热成像深度估计

热成像能够反映场景的温度分布,对热成像进行深度估计,可以恢复出场景的三维温度场,在故障诊断、夜视导航等领域具有重要意义。本文提出一种面向单目热成像深度估计的非参深度采样方法。为了克服热像纹理缺乏、轮廓模糊的缺点,使用了空间金字塔匹配(Spatial Pyramid Matching,SPM)来进行热像的特征分析。首先,基于SPM特征匹配,从数据库中筛选出与待估计深度的热像具有相似场景的候选热像;然后,采用SIFT Flow变形算法对候选热像的深度图进行采样,并将深度信息传递给待估计的热像。实验结果表明,这种方法能够对单目热像进行有效的深度估计,与同类算法相比具有明显优势。     

Automatic foreground extraction from imperfect backgrounds using multi-agent consensus equilibrium

Extracting accurate foreground objects from a scene is an essential step for many video applications. Traditional background subtraction algorithms can generate coarse estimates, but generating high quality masks requires professional softwares with significant human interventions, e.g., providing trimaps or labeling key frames. We propose an automatic foreground extraction method in applications where a static but imperfect background is available. Examples include filming and surveillance where the background can be captured before the objects enter the scene or after they leave the scene. Our proposed method is very robust and produces significantly better estimates than state-of-the-art background subtraction, video segmentation and alpha matting methods. The key innovation of our method is a novel information fusion technique. The fusion framework allows us to integrate the individual strengths of alpha matting, background subtraction and image denoising to produce an overall better estimate. Such integration is particularly important when handling complex scenes with imperfect background. We show how the framework is developed, and how the individual components are built. Extensive experiments and ablation studies are conducted to evaluate the proposed method.     

基于学习的光栅图像噪声抑制方法

基于条纹投影的三维形貌测量广泛应用于工业制造、质量检测、生物医疗、航空航天等领域。然而在高速测量的场景下,由于光栅图像的采集过程曝光时间短,三维重建结果通常会受到较为严重的图像噪声干扰。近年来,深度学习技术在计算机视觉等领域得到了广泛应用,并且取得了巨大的成功。受此启发,提出了一种基于学习的光栅图像噪声抑制方法。首先构建了一个基于U-net的卷积神经网络。其次在训练过程中,构建的神经网络学习从含有噪声的条纹图像到对应高质量包裹相位之间的映射关系。当经过适当训练,该网络可从含有噪声的条纹图像中准确恢复相位信息。实验结果表明:针对离线的快速运动场景三维测量,该方法仅利用一幅光栅图像可恢复高精度的相位信息,且相位精度优于传统的三步相移方法。该方法可为提升运动高速场景三维测量的精度提供切实可靠的解决方案。     

In-camera automation of photographic composition rules.

At the time of image acquisition, professional photographers apply many rules of thumb to improve the composition of their photographs. This paper develops a joint optical-digital processing framework for automating composition rules during image acquisition for photographs with one main subject. Within the framework, we automate three photographic composition rules: repositioning the main subject, making the main subject more prominent, and making objects that merge with the main subject less prominent. The idea is to provide to the user alternate pictures obtained by applying photographic composition rules in addition to the original picture taken by the user. The proposed algorithms do not depend on prior knowledge of the indoor/outdoor setting or scene content. The proposed algorithms are also designed to be amenable to software implementation on fixed-point programmable digital signal processors available in digital still cameras.     

基于多分辨率格网的三维物体识别方法

本文首先提出了一种改进的三维物体表达方法,它将一个三维物体表面网格与其它表面网格的几何关系表示为一个二维矩阵,称为距离角度图.这种表达能够描述任意形态物体,抑制杂乱背景和遮挡,几何意义直观,且适应不同分辨率、非规则的三角格网.然后,以这种表达方法为基础,本文阐述了一种基于多分辨率格网的,由粗到精的三维物体识别方法.它先在场景和模型的低分辨率格网上进行粗匹配以得到模型候选集合,之后在已匹配网格的高分辨率格网邻域上筛选模型候选集合,最后综合考虑多个网格对应的模型候选以得到最终模型候选的确认和验证.这种识别方法具有运算量小,准确可靠等优点,实验证明该方法正确有效.     

基于双目立体视觉的场景分割方法

提出一种基于双目立体视觉的场景分割方法:首先根据双目立体视觉系统提供的左右视图进行三维场景重构,得到场景的几何深度图,同时利用左视图进行RGB颜色空间到CIELab均匀颜色空间的转换以得到颜色信息;然后将颜色与几何信息构造生成六维向量;最后再将六维向量给到聚类算法中进行分割并对分割的伪影进行消除,得到最终的分割结果.对Middlebury数据集样本场景baby 2实验了6种立体视觉算法和3种聚类技术的不同组合进行的场景分割,从实验结果来看,不同的组合应用所提方法都比传统方法具有更好的分割效果.