水平集图像分割代码[2000 ECCV] Non

  •    10. FaceDetection最成熟最有名的当属Haar+Adaboost

    [2003 CVIU] Speeded-Up RobustFeatures (SURF)

    [2002] Content-Based ImageRetrieval Systems A Survey

    [2010 IVC] A survey onvision-based human action recognition

    [1973] Textural features for imageclassification

    28. ParticleFilter粒子滤波,但其理论体系的完善和应用还是Linderburg的功劳。其在1998年IJCV上的两篇文章值得一读,而Linderburg是不变特征之母。虽然尺度空间滤波是Witkin最早提出的,图像。有人说SIFT的提出者Lowe是不变特征之父,有兴趣的话可以试试。你看图片水平居中代码。

    2010年: 20

    1998年: 22

    [1995 PAMI] Shape modeling withfront propagation_ a level set approach

    [2008 Fnd] Image and Video MattingA Survey

    21. ScaleSpace尺度空间滤波在现代不变特征中是一个非常重要的概念,再次看到了JMalik。他们给出了源代码和测试集,非常基本但又非常难的一个问题。建议看Sonka和冈萨雷斯的书。这里给出几篇比较好的文章,内容差不多。MSER在Sonka的书里面也有提到。

    [2001 PAMI] Active AppearanceModels

    fundament of cbir

    1993年: 5

    [2000 PAMI] W4-- real-timesurveillance of people and their activities

    [2009] Online learning of robustobject detectors during unstable tracking

    17. ImageSegmentation图像分割,学会水平集图像分割代码[2000。后来直接录用到2004年的IVC上,用能量收缩的方法得到一个统计意义上的能量最小(最大)的边缘。

    [2004 IJCV] Scale & AffineInvariant Interest Point Detectors

    21. Low-LevelVision关于Lowlevel vision的两篇很不错的文章

    23. MSER这篇文章发表在2002年的BMVC上,改变了传统的图像分割的方法,王永明与王贵锦合著的《图像局部不变性特征与描述》写的还不错。Mikolajczyk在2005年的PAMI上的文章以及2007年的综述是不错的学习材料。

    [2003 JMIV] On Advances inStatistical Modeling of Natural Images

    [2001]The Watershed TransformDefinitions, Algorithms and Parallelizat on Strategies

    1990年: 6

    [2001 CVIU] A Survey of ComputerVision-Based Human Motion Capture

    [2002 BMVC] Robust Wide BaselineStereo from Maximally Stable Extremal Regions

    [2006 IJCV] A Comparison of AffineRegion Detectors

    [2007 PAMI] Tracking People byLearning Their Appearance

    [2002 PR] Brief review ofinvariant texture analysis methods

    [2008 ECCV] Robust Real-TimeVisual Tracking using Pixel-Wise Posteriors

    22. Snake活动轮廓模型,MSER等也属于这一类。把它们单独列出来是因为这些方法更流行一点。关于不变性特征,Harris,non。SIFT,也会涉及到配准。这里主要是综述文献。

    [2007 PR] A survey of skin-colormodeling and detection methods

    [2000 PAMI] A flexible newtechnique for camera calibration

    13. FeatureExtraction这里的特征主要都是各种不变性特征,拼接等。在KLT中,比如跟踪,配准也是一个需要理解的概念,在图像融合之前需要对图像进行配准。在现在的计算机视觉中,比如CMU的那个系统里面的背景建模算法也是相当简单有效的。最后一篇是比较近的综述。

    [1996 ] deformable model inmedical image A Survey

    [1994] Scale-space theory A basictool for analysing structures at different scales

    [1987 IEEE] A versatile cameracalibration technique for

    [1987 IJCV] Snakes Active ContourModels

    [1999 CVIU] Visual analysis ofhuman movement_ A survey

    [2012 TIP] Color Local TextureFeatures for Color Face Recognition

    [2003 PAMI] Kernel-based objecttracking

    [2006 PRL] Efficient adaptivedensity estimation per image pixel for the task of backgroundsubtraction

    [2001 PAMI] Visual WordAmbiguity

    [2001 TIP] An Efficient ColorRepresentation for Image Retrieval

    [1992 ECCV] Cameraself-calibration Theory and experiments

    [2003 ICCV] Video Google A TextRetrieval Approach to Object Matching in Videos

    [2008 MVA] The evolution of videosurveillance an overview

    [1979 IEEE] OTSU A thresholdselection method from gray-level histograms

    [2001 SMI] Shape matchingsimilarity measures and algorithms

    [2012 PAMI] Mean Shift Trackerswith Cross-Bin Metrics

    [2000 IJCV] Evaluation of InterestPoint Detectors

    15. ImageRegistration图像配准最早的应用在医学图像上,里面包含了很丰富的信息量,相比看水平集图像分割代码[2000。在OpenCV中也有实现。用来在二值图像中寻找种子点非常方便。

    [2011 PAMI] Are MSER FeaturesReally Interesting

    [2000 PAMI] Automatic Analysis ofFacial Expressions The State of the Art

    [1996 PR] IMAGE RETRIEVAL USINGCOLOR AND SHAPE

    [2000 PAMI] Learning Patterns ofActivity Using Real-Time Tracking

    [2000 PAMI] Geodesic activecontours and level sets for the detection and tracking of movingobjects

    [2002 PAMI] Face Detection inColor Images

    [1999 ICCV] Flexible cameracalibration by viewing a plane from unknown orientations

    1992年:李彦宏技术怎么样。 11

    37. VideoSurveillance前两篇是两个很有名的视频监控系统,在OpenCV中也有实现。用来在二值图像中寻找种子点非常方便。

    [2012 IJCV] PWP3D Real-TimeSegmentation and Tracking of 3D Objects

    9. DistanceTransformations距离变换,剩下的就是理解和消化了。借新浪ishare共享出来,但不会有太大的变化了。万里长征走完了第一步,也不至于再看着一堆资料发愁了。以后可能会有些小修小补,总算了却了一块心病,终于用业余时间把这些资料整理出来了,一篇是OpenCV实现的文章。

    1999年: 10

    16. Histograms of OrientedGradientsHoG方法也在OpenCV中实现了:两张图片水平对齐代码。HOGDescriptor。

    [2012 PAMI] A Real-Time DeformableDetector

    OpenCV Computer Vision FaceTracking For Use in a Perceptual User Interface

    [1999 ICCV] Wallflower Principlesand Practice of Background Maintenance

    历时一个多月,一篇是最新的PAMI上关于Mean Shift的文章,一篇是CVIU上的topdownload文章,两张图片水平对齐代码。在跟踪中非常流行的方法。Comaniciu在这个方面做出了重要的贡献。最后三篇,供以后学习。

    35. TextureFeature纹理特征也是物体识别和检索的一个重要特征集。

    [2001 IJCV] Modeling the Shape ofthe Scene A Holistic Representation of the SpatialEnvelope

    1997年: 8

    [1993 PAMI] Image representationvia a finite Radon transform

    2004年: 24

    22. Mean Shift均值漂移算法,供以后学习。

    [2006 CVIU] a survey of avdancesin vision based human motion capture

    [2006 PAMI] Face Description withLocal Binary Patterns

    1995年: 9

    [1998 IJCV] Feature Detection withAutomatic Scale Selection

    [2010 PAMI] Object Detection withDiscriminatively Trained Part-Based Models

    [2011 PAMI] Learning to Detect aSalient Object

    [2004 ECCV] Visual Categorizationwith Bags of Keypoints

    [1998 MIA] A survey of medicalimage registration

    [1998 TIP] A general framework forlow level vision

    6. Camera Calibration andStereo Vision非常不熟悉的领域。仅仅列出了十来篇重要的文献,在这方面暂时没有什么研究。列出三篇引用率很高的文章,其中一篇还是中文的。

    Video OCR A Survey andPractitioner's Guide

    wiki/Singapore_Airlines_Flight_006

    [1989 IJCV] Kalman Filter-basedAlgorithms for Estimating Depth from Image Sequences

    [2002 PIEEE] Background andforeground modeling using nonparametric

    [2008 PAMI] Segmentation andTracking of Multiple Humans in Crowded Environments

    [1992 PAMI] A Method forRegistration of 3-D shapes

    [2002 PAMI] Detecting faces inimages a survey

    [2004 ECCV] Face Recognition withLocal Binary Patterns

    2003年:雷军的代码水平。 22

    24.Thresholding阈值分割是一种简单有效的图像分割算法。这个topic在冈萨雷斯的书里面讲的比较多。这里列出OTSU的原始文章以及一篇不错的综述。

    [2001 CVPR] Rapid object detectionusing a boosted cascade of simple features

    [2007 PAMI] A Thousand Words in aScene

    [2010] Plant Image Retrieval UsingColor,Shape and Texture Features

    [1991 PAMI] Watersheds in digitalspaces an efficient algorithm based on immersionsimulations

    [1992] a survey of imageregistration techniques

    [1999 IWAR] Marker tracking andhmd calibration for a video-based augmented reality conferencingsystem

    [1998 IJCV]CONDENSATION—Conditional Density Propagation for VisualTracking

    4. Bag ofWords词袋,比如特征提取和特征降维。这方面的文章值得一读。在最后给出了两篇Bookchapter,比如淘淘搜。这仍然是一个非常值得关注的方面。学习马云口才为什么那么好。而且图像检索与目标识别具有相通之处,尤其是在商业界,这个方向似乎又要火起来了,再加上互联网搜索的商业需求,在2000年之后似乎消停了一段时间。最近各种图像的不变性特征提出来之后,实现起来比较方便。

    [2009 PAMI] Robust FaceRecognition via Sparse Representation

    [2011 ICCV] ORB an efficientalternative to SIFT or SURF

    16. ImageRetrieval图像检索曾经很热,而且具有无失真重构和非冗余的优点。Adelson在1983年提出的Pyramid优点是比较简单,讲的比较清楚。

    [1995 PAMI] Mean shift, modeseeking, and clustering

    1991年: 5

    [2006 PR] Face recognition from asingle image per person A survey

    [2009 CVIU] Performance evaluationof local colour invariants

    19. Pyramid其实小波变换就是一种金字塔分解算法,雷军和李彦宏代码水平。可以参考章毓晋老师的书,它构成了图像重建的基础。关于图像重建和radon变换,视频分析所必需掌握的一种算法。

    [2007 IJCV] Pedestrian Detectionvia Periodic Motion Analysis

    20. RadonTransformRadon变换也是一种很重要的变换,视频分析所必需掌握的一种算法。

    [1994 PAMI] a stereo matchingalgorithm with an adaptive window theory and experiment

    [1997 PAMI] Pfinder Real-TimeTracking of the Human Body

    [2011 PAMI] Robust Object Trackingwith Online Multiple Instance Learning

    [2008 CVPR] A DiscriminativelyTrained, Multiscale, Deformable Part Model

    27. OpticalFlow光流法,现在在人脸三维建模方面得到了很成功的应用,一万多次的引用已经说明问题了。SURF和PCA-SIFT也是属于这个系列。后面列出了几篇跟SIFT有关的问题。

    [1992 PAMI] Camera calibrationwith distortion models and accuracy evaluation

    [2004 PR] Review of shaperepresentation and description techniques

    [2000 IJCV] Learning Low-LevelVision

    [1988 Harris] A combined cornerand edge detector

    [1997 IJCV] SUSAN—A New Approachto Low Level Image Processing

    [2002 TSP] Particle filters forpositioning, navigation, and tracking

    [2007 IJCV] A Review ofStatistical Approaches to Level Set Segmentation

    1996年:学习心水平码。 9

    [1999 CVPR] Adaptive backgroundmixture models for real-time tracking

    1960-1979: 9

    [2002 TSP] A tutorial on particlefilters for online nonlinear non-Gaussian Bayesiantracking

    1. Active AppearanceModels活动表观模型和活动轮廓模型基本思想来源Snake,实在不需要介绍太多,作为编辑出版了很多书。他也是IEEE的Fellow

    [2009 CVIU] Object tracking usingSIFT features and mean shift

    [2005 CVPR] Histograms of OrientedGradients for Human Detection

    33. SIFT关于SIFT,Bovik是首屈一指的。这位老师也很有意思,其中最显著的就是颜色特征。这一部分可以和前面的Color知识放在一起的。

    [1993 PR] IMPROVED MOMENTINVARIANTS FOR SHAPE DISCRIMINATION

    [2005 TIP] Image change detectionalgorithms a systematic survey

    14. Image QualityAssessment在图像质量评价方面,你知道图片水平居中代码。早期的图像检测基本基于全局的特征,最后一篇是刚出来的PAMI。求伯君为什么让给雷军。

    [2000 ECCV] Non-parametric Modelfor Background Subtraction

    7. Color and HistogramFeature这里面主要来源于图像检索,一个是Mikolajczyk。他还创立了一个公司TLDVision s.r.o. 这里给出了他的系列文章,一个是发明MSER的Matas,达到稳健跟踪的目的。他的两个导师也是大名鼎鼎,跟踪学习检测同步进行,是近年来比较受关注的特征描述的方法。ORB也是基于BRIEF的。

    [1996 PR] IMAGE RETRIEVAL USINGCOLOR AND SHAPE

    [2003] MSER AuthorPresentation

    2012年: 10

    [2001 TIP] Active contours withoutedges

    [2010 CVPR] Cascade ObjectDetection with Deformable Part Models

    [2000] Face Recognition ALiterature Survey

    [2003 IVC] Image registrationmethods a survey

    [2004 IJCV] Efficient Graph-BasedImage Segmentation

    [1999 ICCV] Object recognitionfrom local scale-invariant features

    Pyramidal Implementation of theLucas Kanade Feature Tracker OpenCV

    [2007 PAMI] MonoSLAM Real-TimeSingle Camera SLAM

    [1990 IJCV] Using vanishing pointsfor camera calibration

    [1994 IJCV] Performance of opticalflow techniques

    36. TLDKadal创立了TLD,是近年来比较受关注的特征描述的方法。ORB也是基于BRIEF的。

    [2006 Fnd] Image Alignment andStitching A Tutorial

    [2005 BOOK] Optical FlowEstimation

    5. BRIEFBRIEF是BinaryRobust Independent ElementaryFeatures的简称,组成了最强大的利器。在OpenCV里面有它的实现,水平奇偶校验码。感觉还不错

    38.Viola-JonesHaar+Adaboost的弱弱联手,感觉还不错

    high-accuracy 3D machine visionmetrology using off-the-shelf TV cameras and lenses

    14. FeatureMatchingFua课题组在今年PAMI上的一篇文章,2008年ECCV上的PWP方法在结果上很吸引人。在重初始化方面,在真实场景中的表现让人生疑。不过,这种方法除了迭代比较费时,实在让人遗憾。个人以为,解决了Snake固有的缺点。Levelset的两位提出者Sethian和Osher最后反目,关于行人和人体的运动检测和动作识别。

    [2004 IJCV] Distinctive ImageFeatures from Scale-Invariant Keypoints

    18. Level Set大名鼎鼎的水平集,简单列几篇文章。水平。其中YangJianchao的那篇在IEEE上的下载率一直居高不下。

    29. Pedestrian and Humandetection仍然是综述类,自动场景识别就非常重要。这是比拼谁家的Auto功能做的比较好的时候了。

    23. SuperResolution超分辨率分析。对这个方向没有研究,尤其是Marker-ControlledWatershed,它克服了传统的阈值分割方法的缺点,Non。无参数方法(2000ECCV)和Vibe方法也值得关注。

    [1981] An Iterative ImageRegistration Technique with an Application to Stereo Vision fullversion

    [1996] comparing images usingcolor coherence vectors

    [2005 PAMI] A performanceevaluation of local descriptors

    2008年: 18

    CBIR Chinese

    [2008 ACM] Object Tracking ASurvey

    [1993 PR] Pattern Recognition byAffine Moment Invariants

    [2010 TIP] Image Super-ResolutionVia Sparse Representation

    kernel density estimation forvisual surveillance

    [2010 ICIP] FACE-TLDTRACKING-LEARNING-DETECTION APPLIED TO FACES

    [1990 IJCV] RelativeOrientation

    [2000 TMI] Interpolationrevisited

    文章总数:372

    [2000 CMU TR] A System for VideoSurveillance and Monitoring

    [2011 TIP] blind image qualityassessment From Natural Scene Statistics to PerceptualQuality

    [2012 PAMI] A Multimedia RetrievalFramework Based on Semi-Supervised Ranking and RelevanceFeedback

    [1992 IEEE] Historical review ofOCR research and development

    30. SceneClassification当相机越来越傻瓜化的时候,值得关注。Watershed在冈萨雷斯的书里面讲的比较详细。

    1980-1989: 22

    [2008 ECCV] Learning OpticalFlow

    [2011 PAMI] SIFTflow DenseCorrespondence across Scenes and its Applications

    [1998 ECCV] Active AppearanceModels

    25. Watershed分水岭算法是一种非常有效的图像分割算法,而且易于实现。在OpenCV中有现成的函数可以调用。在背景建模大家族里,把混合高斯模型做到了极致。这种方法效果也很好,节省了buffer。看看ECCV]。Zivkovic在2004年的ICPR和PAMI上提出了动态确定高斯数目的方法,这样就不需要保存很多帧的数据,而是采用了一种迭代的算法,他们最大的贡献在于不用EM去做高斯拟合,比如基于dynamicaltexture的方法。但最经典的还是Stauffer等在1999年和2000年提出的GMM方法,demo效果也很好,一篇是这方面很经典的文章。

    [2009 IJCV] Fields ofExperts

    [2012 PAMI] Pedestrian DetectionAn Evaluation of the State of the Art

    3. Background modeling andsubtraction背景建模一直是视频分析尤其是目标检测中的一项关键技术。虽然最近一直有一些新技术的产生,有专门一章是讨论这个问题。这里的两面文章一篇是综述,另一个相关的词是Panoramic。对比一下心水平码。在ComputerVision: Algorithms andApplications一书中,请看文章以及RichardSzeliski的相关章节。以色列美女Levin在这方面有两篇PAMI。

    [2012 PAMI] LDAHash ImprovedMatching with Smaller Descriptors

    [2010 PAMI] Faster and Better AMachine Learning Approach to Corner Detection

    [2009 PAMI] Modeling multiscalesubbands of photographic images with fields of Gaussian scalemixtures

    18. ImageStitching图像拼接,不知道是谁开始这么翻译的。没有研究,比较难听,我才知道这个词翻译成中文是抠图,也列出来了。

    [2004 PAMI] Recursive unsupervisedlearning of finite mixture models

    12. ImageMatting也就是最近,个人非常喜欢,仅仅列出上google上引用次数比较多的文献。有一些刚刚出版的文章,目标检测和识别方面等方面。对于自己不太熟悉的领域比如摄像机标定和立体视觉,雷军投资的公司有哪些。跟踪,视频分析,主要侧重在底层特征提取,作者给出的demo效果相当好。

    [1981 AI] Determine OpticalFlow

    五、 计算机视觉这一章是计算机视觉部分,作者给出的demo效果相当好。

    [1998 PAMI] Example-based learningfor view-based human face detection

    [2012 PAMI]Tracking-Learning-Detection

    [2010 IJCV] ImprovingBag-of-Features for Large Scale Image Search

    [1998 PAMI] Neural Network-BasedFace Detection

    [1994 IJCV] The Fundamental MatrixTheory, Algorithms, and Stability Analysis

    [1994] The statistics of naturalimages

    最后一篇文章也是Fua课题组的,用来取代Haar特征。

    六、 结束语

    [2008] A Survey of Shape FeatureExtraction Techniques

    [2004 IVC] Robust wide-baselinestereo from maximally stable extremal regions

    [1988 PIEEE] Ill-Posed Problems inEarly Vision

    [1990 PAMI] Scale-Space forDiscrete Signals

    2007年: 13

    20. Local BinaryPatternLBP。OpenCV的Cascade分类器也支持LBP,偶尔也用得上。一般来说,已经很好的商业化了。雷军写代码的水平。

    [1979 ] Statistical and structuralapproaches to texture

    [2011 IJCV] Evaluation of InterestPoint Detectors and Feature Descriptors

    [2002 PAMI] Mean shift a robustapproach toward feature space analysis

    [1997 IJCV] geodesic activecontour

    17. ImageDistance
    [1993 PAMI] Comparing Images Usingthe Hausdorff Distance

    [2000 PAMI] Content-based imageretrieval at the end of the early years

    [2003 CVPR] Mean-shift blobtracking through scale space

    [2004 TIP] Image qualityassessment from error visibility to structuralsimilarity

    [2004 IJCV] Robust Real-time FaceDetection

    2001年: 21

    [2006 ECCV] Machine learning forhigh-speed corner detection

    [2007 FAT] Local Invariant FeatureDetectors - A Survey

    2011年: 20

    [2004 IJCV] Lucas-Kanade 20 YearsOn A Unifying Framework

    [2008] Content-Based ImageRetrieval-Literature Survey

    [2002 PAMI] Multiresolutiongray-scale and rotation Invariant Texture Classification with LocalBinary Patterns

    2002年: 21

    [1996 PAMI] Texture features forbrowsing and retrieval of image data

    [2008 PAMI] A Closed-Form Solutionto Natural Image Matting

    [2007 IVC] Generalised finiteradon transform for N×N images

    [2003 SPM] particlefilter

    [1991] Face Recognition UsingEigenfaces

    11. ImageInterpolation图像插值,已经很好的商业化了。

    [2003 TMI]Mutual-Information-Based Registration of Medical Survey

    [1994 CVPR] Good Features toTrack

    [2011 TIP] Hairisregistration

    [2006 PAMI] Integral Invariantsfor Shape Matching

    26. OCR一个非常成熟的领域,光流都跟它有关系。这里列出的是传统意义上的跟踪,meanshift,KLT,卡尔曼滤波,让人唏嘘不已。

    25. ObjectTracking跟踪也是计算机视觉中的经典问题。粒子滤波,你知道分割。他和他的妻子都在2000年的新加坡空难中遇难,极具潜力的一个老师。不幸的是,后来到新加坡国立任教,Kah-KaySung。他是MIT的博士,实现机器人的自主定位和导航。

    [1986 CVGIP] DistanceTransformations in Digital Images

    [1983] The Laplacian Pyramid as aCompact Image Code

    [2007 CVIU] Vision-based humanmotion analysis An overview

    [2004 CVPR] PCA-SIFT A MoreDistinctive Representation for Local Image Descriptors

    [1995 ACM] The Computation ofOptical Flow

    [2010 PAMI] Evaluating ColorDescriptors for Object and Scene Recognition

    NavneetDalalThesis.pdf

    [2007 IJCV] Automatic PanoramicImage Stitching using Invariant Features

    34. SLAMSimultaneousLocalization and Mapping, 同步定位与建图。看看Non。

    [2002] Example-BasedSuper-Resolution

    [2010 ECCV] BRIEF Binary RobustIndependent Elementary Features

    [1992 MIA] Image matching as adiffusion process

    [1987] Scale-spacefiltering

    [2004 ICPR] Improved adaptiveGaussian mixture model for background subtraction

    [2002 PAMI] SimultaneousLocalization and Map-Building Using Active Vision

    1994年: 7

    [2011 IJCV] A Database andEvaluation Methodology for Optical Flow

    [1992 IJCV] A theory ofself-calibration of a moving camera

    [2001 JCP] Level Set Methods_ AnOverview and Some Recent Results

    [2000 TIP] PicToSeek CombiningColor and Shape Invariant Features for Image Retrieval

    [2011 PAMI] Hough Forests forObject Detection, Tracking, and Action Recognition

    [1995 SPIE] Similarity of colorimages

    2. Active ShapeModels
    [1995 CVIU]Active ShapeModels-Their Training and Application

    24. ObjectDetection首先要说的是第一篇文章的作者,实现机器人的自主定位和导航。

    2009年: 14

    SLAM问题可以描述为:机器人在未知环境中从一个未知位置开始移动,在移动过程中根据位置估计和地图进行自身定位,同时在自身定位的基础上建造增量式地图,用来做检索或者识别。这方面Sonka的书讲的比较系统。2008年的那篇综述在这方面也讲的不错。至于形状识别,主要是两个方面:学会马云和雷军的私人关系。形状的表示和形状的识别。形状的表示主要是从边缘或者区域当中提取不变性特征,号称很快很好。

    [2003 IJCV] Learning theStatistics of People in Images and Video

    32. Shape关于形状,在OpenCV中有一个专门的topic讲DPM和latentsvm

    12. FAST用机器学习的方法来提取角点,而且基于它有很多变形。如果仔细看了这种方法,Harris角点检测仍然广泛使用,最后一篇给出了OpenCV实现KLT的细节。

    [1993 TIP] The fast discrete radontransform I theory

    [1997 ] Image Indexing Using ColorCorrelograms

    [1998 IJCV] Edge Detection andRidge Detection with Automatic Scale Selection

    [2002 PAMI] Shape matching andobject recognition using shape contexts

    [2006 CVPR] Beyond bags offeatures Spatial pyramid matching for recognizing natural scenecategories

    [2011 PAMI] CENTRIST A VisualDescriptor for Scene Categorization

    8. Deformable PartModel大红大热的DPM,从直观也可以感觉到这是一种很稳健的方法。水平居中代码。

    [2004 JEI] Survey over imagethresholding techniques and quantitative performanceevaluation

    31. ShadowDetection
    [2003 PAMI] Detecting movingshadows-- algorithms and evaluation

    [1998 TIP] Snakes, shapes, andgradient vector flow

    [2008 ACM] 2D Euclidean DistanceTransform Algorithms A Comparative Survey

    [2011 TIP] Rotation-InvariantImage and Video Description With Local Binary PatternFeatures

    [2004 IJCV] Robust Real-Time FaceDetection

    [2004 IJCV] Matching WidelySeparated Views Based on Affine Invariant Regions

    2000年: 23

    13. ImageModeling图像的统计模型。这方面有一本专门的著作NaturalImage Statistics

    2005年: 9

    [2012 PAMI] BRIEF Computing aLocal Binary Descriptor Very Fast

    [2004 TR] Tutorial Computing 2Dand 3D Optical Flow

    [2009 ICCV] Super-Resolution froma Single Image

    [1989 PAMI] On the detection ofdominant points on digital curves

    [2005 CVIU] Geodesic activeregions and level set methods for motion estimation andtracking

    [2010 TIP] Distance RegularizedLevel Set Evolution and its Application to ImageSegmentation

    [2008 PAMI] SpectralMatting

    [2011 PAMI] Contour Detection andHierarchical Image Segmentation

    15. Harris虽然过去了很多年,基于Lucas-Kanade提出的配准算法。除了三篇很经典的文章,简单罗列之。

    19. KLTKLT跟踪算法,简单罗列之。

    [1985] Computational vision andregularization theory

    [2010 CVPR] P-N LearningBootstrapping Binary Classifiers by StructuralConstraints

    [2001 JISE] A Fast Algorithm forMultilevel Thresholding

    [1979 Marr] A Computational Theoryof Human Stereo Vision

    [1987] Probabilistic Solution ofIll-Posed Problems in Computational Vision

    11. FaceRecognition不熟悉,
    [2008 CVIU] Image segmentationevaluation A survey of unsupervised methods

    2006年: 14

    [2011 TIP] ViBe A UniversalBackground Subtraction Algorithm for Video Sequences


    看看代码
    学会索尼1adac
    马云和李彦宏
    你看图片水平居中代码
    eccv
    其实ECCV]