Removing Shading Distortions in Camera-based Document Images Using Inpainting and Surface Fitting With Radial Basis Functions
|
|
- Jasmin Tucker
- 7 years ago
- Views:
Transcription
1 Removing Shading Distortions in Camera-based Document Images Using Inpainting and Surface Fitting With Radial Basis Functions Li Zhang Andy M. Yip Chew Lim Tan School of Computing, 3 Science Drive 2, National University of Singapore Department of Mathematics, 2 Science Drive 2, National University of Singapore {dcszl,andyyip,dcstcl}@nus.edu.sg Abstract Shading distortions are often perceived in geometrically distorted document images due to the change of surface normal with respect to the illumination direction. Such distortions are undesirable because they hamper OCR performance tremendously even when the geometric distortions are corrected. In this paper, we propose an effective method that removes shading distortions in images of documents with various geometric shapes based on the notion of intrinsic images. We first try to derive the shading image using an inpainting technique with an automatic mask generation routine and then apply a surface fitting procedure with radial basis functions to remove pepper noises in the inpainted image and return a smooth shading image. Once the shading image is extracted, the reflectance image can be obtained automatically. Experiments on a wide range of distorted document images demonstrate a robust performance. Moreover, we also show its potential applications to the restoration of historical handwritten documents. 1. Introduction The popularity of current hand held digital devices such as digital cameras, cellphones and PDAs has made camera imaging a convenient way of recording information. With such a camera-enabled device, people can snap photos of documents whenever and wherever needed as a way of daily notes taking. However, this also gives rise to many distorted images especially when the imaging environment is uncontrollable. One of such distortions is shading including shadows. Strictly speaking, shading is the variation in luminance caused by a change of surface normal with respect to the illumination direction while shadow refers to the variation caused by occlusions of the light source. In particular, when capturing documents of non-planar geometric shapes, we often receive images containing both geometric and shading distortions. These create great challenges for current OCR systems to identify the words correctly and in the right sequence. To obtain a good recognition rate, it is necessary to correct both distortions to a certain extent. Brown and Tsoi propose a boundary interpolation method to correct both distortions on images of warped art materials [3]. The method produces good results for a variety of geometric warpings but restricted to iso-parametric folding lines. Since the shape of the warped surface is not required, the uniform parametrization needs to be guided by a checkerboard pattern placed beneath the document. Furthermore, image boundaries must be present and an unobstructed white border needs to be enforced for the estimation of the shading. These conditions are often hard to satisfy when people just take snapshots for convenience. Sun et al. present a system to restore both geometric and photometric artifacts of arbitrarily distorted documents [11]. This system requires a special 3D scanning setup to acquire the depth map of the warped surface and it handles mainly nonsmooth shadings caused by folds. On the other hand, methods have been proposed to separate reflectance and illumination images based on the notion of intrinsic images [1], which defines an image as composed of a reflectance component and a shading component. The illumination image here includes both shading and shadow. Color information has been exploited to separate reflectance from shading based on the observation that shading is almost exclusively defined by luminance while reflectance is defined by both luminance and color [9]. Funt et al. [7] propose a method to recover shading from color images by removing reflectance component based on associated abrupt chromaticity changes. In other word, they use the fact that the change of reflectance is usually caused by a change in color. Similarly, Tappen et al. [12] introduce another method to recover shading and reflectance images using both color information and a classifier trained to recognize local gray-scale patterns to distinguish derivatives causedbyreflectance changes from those caused by shading. The intrinsic images are recovered from its derivatives using the same method as introduced by Weiss [14]. In both
2 methods, diffuse surfaces are assumed and the thresholding process can potentially flatten out discontinuous geometric features that may appear in the shading image. Toro et al. [13] describe an approach that addresses both diffuse and specular reflections with a known illumination direction. Despite all these efforts in deriving the intrinsic images, there is no single exact solution because the decomposition of the intensity image into its two intrinsic components is theoretically not unique. In view of the daily snapshots of document images containing mainly text and graphics information, we propose a simple yet effective method that extracts the reflectance image by removing the shading distortions including both geometrically caused shadings and cast shadows. Assuming that a given document has a constant colored background, the reflectance image will contain only the printed texts/graphics which indicate a color change. Our objective is to derive this reflectance image so as to improve the document s visual appearance and the OCR performance. To do this, we first extract the shading image through an inpainting technique followed by a surface fitting process when appropriate. The inpainting mask is obtained using an edge-based method followed by some morphological operations. Once the shading image is obtained, the reflectance image can be derived easily based on the notion of intrinsic images. Experiments on various document images demonstrate a robust performance. In addition, we also show that this method can be used to clean up historical handwritten document images with stains and patch noises. 2. A General Work Flow Figure 1 illustrates a detailed work flow of the proposed method. Given a distorted document image, we first extract an inpainting mask, which masks the text/graphics contents that cause a reflectance change. This is done using an edge-based method followed by a morphological operation. Next, a harmonic or Total Variation (TV) inpainting technique is applied to the original image to remove the printed contents using the extracted mask. If the mask does not fully cover the printed contents, the inpainted image may contain scattered pepper noises due to unremoved ink. This can be further refined through an iterative mask enhancement process. Alternatively, if the shading is smooth, a surface fitting scheme can be exploited to eliminate the noises and produce a smooth shading image. Once the shading image is extracted, the reflectance image can be easily derived based on the notion of intrinsic images. 3. Shading Extraction using Inpainting Assuming the given document image has a uniformcolored background such as the normal printed notes, plain Figure 1. Work flow of the restoration method. book pages, etc., an effective cue for differentiating shading from reflectance is the printed regions. It has been observed that luminance variations accompanied by color variations are usually variations in reflectance while luminance variations unaccompanied by color variations are variations in illumination [9]. Therefore, the printed text regions essentially imply the reflectance changes. If we can remove the luminance variations caused by the colored text, we will be left with pure shading variations. Therefore, the first step is to identify the text/graphics locations and remove all the colors that have high contrast to the background Automatic Mask Generation Text localization has been a widely researched area either on document images or digital videos. The techniques can be broadly classified as component-based [6, 10] or texture-based [16, 8]. The component-based methods usually try to analyze the geometrical arrangements of edges or uniform colored components of the characters. The texturebased methods utilize the texture characteristics of text lines to extract the text. Here we are interested in not only texts but also graphics. Whatever that may induce a reflectance change is within our consideration. Therefore, we make use of an edge-based method that essentially identifies pixels that are of high contrast to the background. Next, morphological operations are applied to the edge-detected image, which generates a mask for the printed contents. The detailed procedures are as follows: 1) Convert color images into gray-scale. This can be done by picking the luminance component of a color model such as the V-component of the HSV model or the I-component of the HSI model; 2) Detect edges using canny edge detector. Post-processings such as non-maximum suppression and streaking elimination are also applied for better results; 3) Perform morphological dilation followed by closing. The size of the structuring element can be tuned manually or adjusted automatically based on an estimated average character height when applicable.
3 3.2. Harmonic/TV Inpainting Once the mask of the printed regions is generated, an inpaintingtechniquecan be used tofill up the masked regions based on the neighboring background pixels. This is essentially to recover the shading in the printed regions based on the assumption that the local variation of shading is small. Digital inpainting was pioneered by Bertalmio et al. [2] and has since been applied to a variety of image processing applications. Here we use it as a way of recovering the shading. In particular, we look at two non-texture variational inpainting models, harmonic and TV inpainting [5]. Mathematically, inpainting can be considered as a local interpolation problem: Given an image I 0 with a hole H inside, we want to find an image I that matches I 0 outside the hole and has consistent information inside the hole. To do this, we try to find I that minimizes the following energy in a continuous domain : E(I) = χ (I I 0 ) 2 dx + λ I 2 dx (1) where λ>0 is a smoothness parameter and χ denotes the characteristic function: { 1, x \ H χ(x) = (2) 0, otherwise To minimize the energy in Eq. 1, we solve the Euler- Lagrange equation: E I =2[χ (I I 0) λδi] =0 (3) By using a gradient-descent method and a discretization using finite difference, we obtain the iterative update formula: I n+1 i,j ( λ = Ii,j n +Δt h 2 (In i+1,j + In i 1,j + In i,j+1 + In i,j 1 ) 4Ii,j) n χ i,j (Ii,j n I 0i,j ) (4) where h is the grid size and the smoothness parameter λ is chosen through trial and error. The time step Δt can be any small constant that makes the iteration stable. We noticed that the harmonic inpainting constructs a smooth solution which may cause problems when the text/graphics at image boundaries are masked out or when interior edges are occluded due to the overlaid text. This can be remedied by using TV inpainting. Instead of using a penalty term I 2 dx in Eq. 1, which is infinite for discontinuous functions, we use I dx instead, which allows discontinuous functions as minimizers. The energy function now becomes: E(I) = χ (I I 0 ) 2 dx + λ I dx (5) where λ =2σ 2 /ν. A minimizer for this energy function can be computed using a similar scheme as for harmonic inpainting. Note that both harmonic and TV inpainting are essentially local models, in which the inpainting is mainly determined by the existing information I 0 in the vicinity of the inpainted domain H. Moreover, Eq. 1 has a built-in denoising capacity so that it is robust to noise. The main difference is that harmonic inpainting builds very smooth solutions and thus does not cope well with edges, while TV inpainting is able to restore narrow broken smooth edges which often exist in document images due to overlaid texts Surface Fitting with RBF With the mask generated, the inpainting process removes all the masked text/graphics and returns a first-hand shading image. However, the result is often not ideal due to the errors in the extracted mask. For example, some unmasked printed pixels will be considered as background and therefore cause pepper noises in the inpainted image. One way to solve this problem is to iteratively improve the mask until no sharp edges are identified. Alternatively, we can remove the pepper noises by using a surface fitting algorithm with radial basis functions (RBF) [4]. This is especially useful when the shading needs to be smooth for further surface reconstruction tasks. Typically, given a set of 3D points {(x i,f(x i )), i=1, 2,,m} where x i is the x-y coordinate and f(x i ) is the z coordinate, a fitted surface can be expressed as: n g(x) = α j h(x y j ) (6) j=1 where {y j,j=1, 2,,n} is a set of selected collocation points and h(x) is the radial basis function. The number of collocation points is selected based on the dimension of the image, e.g for the image in Figure 2. The goal is to find the coefficients α j that minimizes the least square error defined as: { m } e = min (g(x i ) f(x i )) 2 (7) α 1,,α n i=1 with optional boundary conditions.various kernel functions of different smoothness can be used. Here we use Multiquadrics: h(x) = x 2 + c 2,wherec is a constant with c =10in our experiments. The advantages of using RBF fitting are: 1) It gives explicit formula for derivatives which are more accurate and less noisy than finite difference; 2) It is also easy to incorporate various types of boundary conditions; 3) Unlike polynomial fitting, RBF is more flexible and can be used to fit more complicated surfaces. Finally, Figure 2 shows an example of how surface fitting helps to extract smooth shading images.
4 (a) (b) (c) comparing to 86.7% on the original distorted images. Besides daily snapshots of printed documents, we also evaluated our method on digitized images of historical handwritten documents. These documents contain substantial noises due to the deterioration of the materials and nonuniform lighting. Figure 4 shows that our method can help clean up the noises and return a better image for further Document Image Analysis (DIA) tasks. (d) (e) (f) Figure 2. (a) Image of an arbitrarily warped document page; (b) Extracted inpainting mask; (c) One-pass inpainted image; (d) Shading image using RBF fitting; (e) Fitted 3D surface of the shading image; (f) Extracted reflectance image. 4. Deriving the Reflectance Image Once the shading image is extracted, it is easy to derive the reflectance image based on the notion of intrinsic images. For Lambertian surfaces, the intensity image is the product of the shading image and the reflectance image [1]. Consider the luminance component of the HSV model, we have I = I s I r. Now given the shading image I s,thereflectance image I r can be computed as: I r = e log I log Is. 5. Experimental Results We have evaluated the proposed method on a set of images captured using both normal digital cameras and cellphone cameras. Figure 3(a 1 ) shows a multi-folded paper with printed characters taken in a complex lighting environment. Figure 3(a 3 ) shows that the folded edges are well restored using the TV inpainting algorithm. Figure 3(b 1 ) and (b 4 ) shows a warped map image and its restored reflectance image, respectively. This demonstrates that our method can also deal with graphical documents as long as the background is of constant color. In addition, (a 5 ) and (b 5 ) show the images after geometric restoration which is done independently from the current work as reported in [15]. Next, Figure 3(c 1 ) shows an image taken using a cellphone camera with the phone s shadow on it and (c 4 ) shows the extracted reflectance image with the shadow removed. Lastly, Figure 3(d 1 ) is an image of a pure text document taken using cellphone camera with non-uniform lightings. A set of such text images with different lightings and shadows are used for conducting OCR experiments. For a total of 2,600 words out of 30 document images, we obtained an average word precision of 98.8% on the restored images (a) (b) Figure 4. (a) Noisy historical handwritten documents; (b) Noise cleaned images. Due to the uncontrolled imaging environment, the shading may contain arbitrary illumination variations or cast shadows. Therefore, it does not follow any exact illumination model. Our method here provides a way of estimating the shading and is specifically designed for document images. In some cases where the documents contain images of non-uniform colors such as embedded figures, we can create a mask that covers the whole image and then apply the inpainting algorithm. In addition, this is a standalone shading removal method, which can handle either flat images or geometrically distorted images and does not rely on a shape recovery process. 6. Conclusions In this paper, we propose a method that removes various shading artifacts from distorted document images and recovers the reflectance images for better visualization and further DIA tasks. The main idea is to use the notion of intrinsic images to separate the shading and reflectance images, in which the shading image is extracted based on an inpainting technique followed by a surface fitting procedure to smooth out the noises. Experiments have shown encouraging results and its potential applications to the cleanup of noisy historical documents. Further studies will be carried out to improve the shading extraction algorithm for documents with non-uniform colored background and also those with figures sitting across a folded edge. 7. Acknowledgment This research is supported by A*STAR grant and NUS URC grant R
5 (a 1 ) (a 2 ) (a 3 ) (a 4 ) (a 5 ) (b 1 ) (b 2 ) (b 3 ) (b 4 ) (b 5 ) (c 1 ) (c 2 ) (c 3 ) (c 4 ) (d 1 ) (d 2 ) (d 3 ) (d 4 ) Figure 3. (a 1 )(b 1 )(c 1 )(d 1 ) Original distorted image; (a 2 )(b 2 )(c 2 )(d 2 ) Extracted inpainting mask; (a 3 )(b 3 )(c 3 )(d 3 ) Extracted shading image; (a 4 )(b 4 )(c 4 )(d 4 ) Restored reflectance image; (a 5 )(b 5 ) Geometrically restored image. References [1] H. Barrow and J. Tenenbaum. Recovering intrinsic scene characteristics from images. Computer Vision Systems, pages 3 26, Academic Press, New York, [2] M. Bertalmio, G. Sapiro, C. Ballester, and V. Caselles. Image inpainting. SIGGRAPH 2000, pages , [3] M. S. Brown and Y. C. Tsoi. Geometric and shading correction for images of printed materials using boundary. IEEE Trans. on Image Processing, 15(6): , Jun [4] J. C. Carr, R. K. Beatson, B. C. McCallum, W. R. Fright, T. McLennan, and T. J. Mitchell. Smooth surface reconstruction from noisy range data. Graphite 2003, pages , [5] T. F. Chan and J. H. Shen. Mathematical models for local nontexture inpaintings. SIAM Journal on Applied Mathematics, 62(3): , [6] P. Clark and M. Mirmehdi. Recognizing text in real scenes. Int l Journal on Document Analysis and Recognition, 4: , [7] B. V. Funt, M. S. Drew, and M. Brockington. Recovering shading from color images. 2nd European Conference on Computer Vision, pages , May [8] H. Li, D. Doermann, and O. Kia. Automatic text detection and tracking in digital video. IEEE Trans. on Image Processing, 9(1): , [9] A. Olmos and F. A. A. Kingdom. A biologically inspired algorithm for the recovery of shading and reflectance images. Perception, 33(12): , [10] M. Pietikainen and O. Okun. Edge-based method for text detection from complex document images. Sixth Int l Conf. on Document Analysis and Recognition, pages , [11] M. X. Sun, R. G. Yang, L. Yun, G. Landon, B. Seales, and M. S. Brown. Geometric and photometric restoration of distorted documents. IEEE Int l Conf. on Computer Vision, 2: , Oct [12] M. F. Tappen, W. T. Freeman, and E. H. Adelson. Recovering intrinsic images from a single image. Pattern Analysis and Machine Intelligence, 27(9): , [13] J. Toro, D. Ziou, and M. F. Auclair-Fortier. Recovering the shading image under known illumination. 1st Canadian Conf. on Computer and Robot Vision, pages 92 96, [14] Y. Weiss. Deriving intrinsic images from image sequences. IEEE Int l Conf. on Computer Vision, 2:68 75, [15] L. Zhang, A. M. Yip, and C. L. Tan. Shape from shading based on lax-friedrichs fast sweeping and regularization techniques with applications to document image restoration. Computer Vision and Pattern Recognition, [16] Y. Zhong, H. Zhang, and A. K. Jain. Automatic caption localization in compressed video. IEEE Trans. on Pattern Analysis and Machine Intelligence, 22(4): , 2000.
Highlight Removal by Illumination-Constrained Inpainting
Highlight Removal by Illumination-Constrained Inpainting Ping Tan Stephen Lin Long Quan Heung-Yeung Shum Microsoft Research, Asia Hong Kong University of Science and Technology Abstract We present a single-image
More informationCanny Edge Detection
Canny Edge Detection 09gr820 March 23, 2009 1 Introduction The purpose of edge detection in general is to significantly reduce the amount of data in an image, while preserving the structural properties
More informationBildverarbeitung und Mustererkennung Image Processing and Pattern Recognition
Bildverarbeitung und Mustererkennung Image Processing and Pattern Recognition 1. Image Pre-Processing - Pixel Brightness Transformation - Geometric Transformation - Image Denoising 1 1. Image Pre-Processing
More informationEECS 556 Image Processing W 09. Interpolation. Interpolation techniques B splines
EECS 556 Image Processing W 09 Interpolation Interpolation techniques B splines What is image processing? Image processing is the application of 2D signal processing methods to images Image representation
More informationColorado School of Mines Computer Vision Professor William Hoff
Professor William Hoff Dept of Electrical Engineering &Computer Science http://inside.mines.edu/~whoff/ 1 Introduction to 2 What is? A process that produces from images of the external world a description
More informationAnalecta Vol. 8, No. 2 ISSN 2064-7964
EXPERIMENTAL APPLICATIONS OF ARTIFICIAL NEURAL NETWORKS IN ENGINEERING PROCESSING SYSTEM S. Dadvandipour Institute of Information Engineering, University of Miskolc, Egyetemváros, 3515, Miskolc, Hungary,
More informationA Dynamic Approach to Extract Texts and Captions from Videos
Available Online at www.ijcsmc.com International Journal of Computer Science and Mobile Computing A Monthly Journal of Computer Science and Information Technology IJCSMC, Vol. 3, Issue. 4, April 2014,
More informationEdge detection. (Trucco, Chapt 4 AND Jain et al., Chapt 5) -Edges are significant local changes of intensity in an image.
Edge detection (Trucco, Chapt 4 AND Jain et al., Chapt 5) Definition of edges -Edges are significant local changes of intensity in an image. -Edges typically occur on the boundary between two different
More informationVideo OCR for Sport Video Annotation and Retrieval
Video OCR for Sport Video Annotation and Retrieval Datong Chen, Kim Shearer and Hervé Bourlard, Fellow, IEEE Dalle Molle Institute for Perceptual Artificial Intelligence Rue du Simplon 4 CH-190 Martigny
More informationLighting Estimation in Indoor Environments from Low-Quality Images
Lighting Estimation in Indoor Environments from Low-Quality Images Natalia Neverova, Damien Muselet, Alain Trémeau Laboratoire Hubert Curien UMR CNRS 5516, University Jean Monnet, Rue du Professeur Benoît
More informationFeature Tracking and Optical Flow
02/09/12 Feature Tracking and Optical Flow Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem Many slides adapted from Lana Lazebnik, Silvio Saverse, who in turn adapted slides from Steve
More informationVECTORAL IMAGING THE NEW DIRECTION IN AUTOMATED OPTICAL INSPECTION
VECTORAL IMAGING THE NEW DIRECTION IN AUTOMATED OPTICAL INSPECTION Mark J. Norris Vision Inspection Technology, LLC Haverhill, MA mnorris@vitechnology.com ABSTRACT Traditional methods of identifying and
More informationA Prototype For Eye-Gaze Corrected
A Prototype For Eye-Gaze Corrected Video Chat on Graphics Hardware Maarten Dumont, Steven Maesen, Sammy Rogmans and Philippe Bekaert Introduction Traditional webcam video chat: No eye contact. No extensive
More informationDetection of the Single Image from DIBR Based on 3D Warping Trace and Edge Matching
Journal of Computer and Communications, 2014, 2, 43-50 Published Online March 2014 in SciRes. http://www.scirp.org/journal/jcc http://dx.doi.org/10.4236/jcc.2014.24007 Detection of the Single Image from
More informationAutomatic 3D Reconstruction via Object Detection and 3D Transformable Model Matching CS 269 Class Project Report
Automatic 3D Reconstruction via Object Detection and 3D Transformable Model Matching CS 69 Class Project Report Junhua Mao and Lunbo Xu University of California, Los Angeles mjhustc@ucla.edu and lunbo
More informationA New Image Edge Detection Method using Quality-based Clustering. Bijay Neupane Zeyar Aung Wei Lee Woon. Technical Report DNA #2012-01.
A New Image Edge Detection Method using Quality-based Clustering Bijay Neupane Zeyar Aung Wei Lee Woon Technical Report DNA #2012-01 April 2012 Data & Network Analytics Research Group (DNA) Computing and
More informationVision based Vehicle Tracking using a high angle camera
Vision based Vehicle Tracking using a high angle camera Raúl Ignacio Ramos García Dule Shu gramos@clemson.edu dshu@clemson.edu Abstract A vehicle tracking and grouping algorithm is presented in this work
More informationA PHOTOGRAMMETRIC APPRAOCH FOR AUTOMATIC TRAFFIC ASSESSMENT USING CONVENTIONAL CCTV CAMERA
A PHOTOGRAMMETRIC APPRAOCH FOR AUTOMATIC TRAFFIC ASSESSMENT USING CONVENTIONAL CCTV CAMERA N. Zarrinpanjeh a, F. Dadrassjavan b, H. Fattahi c * a Islamic Azad University of Qazvin - nzarrin@qiau.ac.ir
More informationLOCAL SURFACE PATCH BASED TIME ATTENDANCE SYSTEM USING FACE. indhubatchvsa@gmail.com
LOCAL SURFACE PATCH BASED TIME ATTENDANCE SYSTEM USING FACE 1 S.Manikandan, 2 S.Abirami, 2 R.Indumathi, 2 R.Nandhini, 2 T.Nanthini 1 Assistant Professor, VSA group of institution, Salem. 2 BE(ECE), VSA
More informationSSIM Technique for Comparison of Images
SSIM Technique for Comparison of Images Anil Wadhokar 1, Krupanshu Sakharikar 2, Sunil Wadhokar 3, Geeta Salunke 4 P.G. Student, Department of E&TC, GSMCOE Engineering College, Pune, Maharashtra, India
More informationColour Image Segmentation Technique for Screen Printing
60 R.U. Hewage and D.U.J. Sonnadara Department of Physics, University of Colombo, Sri Lanka ABSTRACT Screen-printing is an industry with a large number of applications ranging from printing mobile phone
More informationThe Heat Equation. Lectures INF2320 p. 1/88
The Heat Equation Lectures INF232 p. 1/88 Lectures INF232 p. 2/88 The Heat Equation We study the heat equation: u t = u xx for x (,1), t >, (1) u(,t) = u(1,t) = for t >, (2) u(x,) = f(x) for x (,1), (3)
More informationNumerical Methods For Image Restoration
Numerical Methods For Image Restoration CIRAM Alessandro Lanza University of Bologna, Italy Faculty of Engineering CIRAM Outline 1. Image Restoration as an inverse problem 2. Image degradation models:
More informationQUALITY TESTING OF WATER PUMP PULLEY USING IMAGE PROCESSING
QUALITY TESTING OF WATER PUMP PULLEY USING IMAGE PROCESSING MRS. A H. TIRMARE 1, MS.R.N.KULKARNI 2, MR. A R. BHOSALE 3 MR. C.S. MORE 4 MR.A.G.NIMBALKAR 5 1, 2 Assistant professor Bharati Vidyapeeth s college
More informationObject Tracking System Using Motion Detection
Object Tracking System Using Motion Detection Harsha K. Ingle*, Prof. Dr. D.S. Bormane** *Department of Electronics and Telecommunication, Pune University, Pune, India Email: harshaingle@gmail.com **Department
More informationSimultaneous Gamma Correction and Registration in the Frequency Domain
Simultaneous Gamma Correction and Registration in the Frequency Domain Alexander Wong a28wong@uwaterloo.ca William Bishop wdbishop@uwaterloo.ca Department of Electrical and Computer Engineering University
More informationLow-resolution Character Recognition by Video-based Super-resolution
2009 10th International Conference on Document Analysis and Recognition Low-resolution Character Recognition by Video-based Super-resolution Ataru Ohkura 1, Daisuke Deguchi 1, Tomokazu Takahashi 2, Ichiro
More informationComparison of different image compression formats. ECE 533 Project Report Paula Aguilera
Comparison of different image compression formats ECE 533 Project Report Paula Aguilera Introduction: Images are very important documents nowadays; to work with them in some applications they need to be
More informationREAL TIME TRAFFIC LIGHT CONTROL USING IMAGE PROCESSING
REAL TIME TRAFFIC LIGHT CONTROL USING IMAGE PROCESSING Ms.PALLAVI CHOUDEKAR Ajay Kumar Garg Engineering College, Department of electrical and electronics Ms.SAYANTI BANERJEE Ajay Kumar Garg Engineering
More informationComputer Graphics. Geometric Modeling. Page 1. Copyright Gotsman, Elber, Barequet, Karni, Sheffer Computer Science - Technion. An Example.
An Example 2 3 4 Outline Objective: Develop methods and algorithms to mathematically model shape of real world objects Categories: Wire-Frame Representation Object is represented as as a set of points
More informationSubspace Analysis and Optimization for AAM Based Face Alignment
Subspace Analysis and Optimization for AAM Based Face Alignment Ming Zhao Chun Chen College of Computer Science Zhejiang University Hangzhou, 310027, P.R.China zhaoming1999@zju.edu.cn Stan Z. Li Microsoft
More informationVariational approach to restore point-like and curve-like singularities in imaging
Variational approach to restore point-like and curve-like singularities in imaging Daniele Graziani joint work with Gilles Aubert and Laure Blanc-Féraud Roma 12/06/2012 Daniele Graziani (Roma) 12/06/2012
More informationConvolution. 1D Formula: 2D Formula: Example on the web: http://www.jhu.edu/~signals/convolve/
Basic Filters (7) Convolution/correlation/Linear filtering Gaussian filters Smoothing and noise reduction First derivatives of Gaussian Second derivative of Gaussian: Laplacian Oriented Gaussian filters
More informationCircle Object Recognition Based on Monocular Vision for Home Security Robot
Journal of Applied Science and Engineering, Vol. 16, No. 3, pp. 261 268 (2013) DOI: 10.6180/jase.2013.16.3.05 Circle Object Recognition Based on Monocular Vision for Home Security Robot Shih-An Li, Ching-Chang
More information3D Scanner using Line Laser. 1. Introduction. 2. Theory
. Introduction 3D Scanner using Line Laser Di Lu Electrical, Computer, and Systems Engineering Rensselaer Polytechnic Institute The goal of 3D reconstruction is to recover the 3D properties of a geometric
More informationHSI BASED COLOUR IMAGE EQUALIZATION USING ITERATIVE n th ROOT AND n th POWER
HSI BASED COLOUR IMAGE EQUALIZATION USING ITERATIVE n th ROOT AND n th POWER Gholamreza Anbarjafari icv Group, IMS Lab, Institute of Technology, University of Tartu, Tartu 50411, Estonia sjafari@ut.ee
More informationA Method of Caption Detection in News Video
3rd International Conference on Multimedia Technology(ICMT 3) A Method of Caption Detection in News Video He HUANG, Ping SHI Abstract. News video is one of the most important media for people to get information.
More informationFast and Enhanced Algorithm for Exemplar Based Image Inpainting
Fast and Enhanced Algorithm for Exemplar Based Image Inpainting Anupam Information Technology IIIT Allahabad, India anupam@iiita.ac.in Pulkit Goyal Information Technology IIIT Allahabad, India pulkit110@gmail.com
More informationThe Scientific Data Mining Process
Chapter 4 The Scientific Data Mining Process When I use a word, Humpty Dumpty said, in rather a scornful tone, it means just what I choose it to mean neither more nor less. Lewis Carroll [87, p. 214] In
More informationAdobe Marketing Cloud Sharpening images in Scene7 Publishing System and on Image Server
Adobe Marketing Cloud Sharpening images in Scene7 Publishing System and on Image Server Contents Contact and Legal Information...3 About image sharpening...4 Adding an image preset to save frequently used
More informationA Study of Automatic License Plate Recognition Algorithms and Techniques
A Study of Automatic License Plate Recognition Algorithms and Techniques Nima Asadi Intelligent Embedded Systems Mälardalen University Västerås, Sweden nai10001@student.mdh.se ABSTRACT One of the most
More informationDetermining optimal window size for texture feature extraction methods
IX Spanish Symposium on Pattern Recognition and Image Analysis, Castellon, Spain, May 2001, vol.2, 237-242, ISBN: 84-8021-351-5. Determining optimal window size for texture feature extraction methods Domènec
More informationComputer Applications in Textile Engineering. Computer Applications in Textile Engineering
3. Computer Graphics Sungmin Kim http://latam.jnu.ac.kr Computer Graphics Definition Introduction Research field related to the activities that includes graphics as input and output Importance Interactive
More informationA Short Introduction to Computer Graphics
A Short Introduction to Computer Graphics Frédo Durand MIT Laboratory for Computer Science 1 Introduction Chapter I: Basics Although computer graphics is a vast field that encompasses almost any graphical
More informationAn Approach for Utility Pole Recognition in Real Conditions
6th Pacific-Rim Symposium on Image and Video Technology 1st PSIVT Workshop on Quality Assessment and Control by Image and Video Analysis An Approach for Utility Pole Recognition in Real Conditions Barranco
More informationRobust and Efficient Implicit Surface Reconstruction for Point Clouds Based on Convexified Image Segmentation
Noname manuscript No. (will be inserted by the editor) Robust and Efficient Implicit Surface Reconstruction for Point Clouds Based on Convexified Image Segmentation Jian Liang Frederick Park Hongkai Zhao
More informationHigh Quality Image Magnification using Cross-Scale Self-Similarity
High Quality Image Magnification using Cross-Scale Self-Similarity André Gooßen 1, Arne Ehlers 1, Thomas Pralow 2, Rolf-Rainer Grigat 1 1 Vision Systems, Hamburg University of Technology, D-21079 Hamburg
More informationSegmentation of building models from dense 3D point-clouds
Segmentation of building models from dense 3D point-clouds Joachim Bauer, Konrad Karner, Konrad Schindler, Andreas Klaus, Christopher Zach VRVis Research Center for Virtual Reality and Visualization, Institute
More informationImplementation of Canny Edge Detector of color images on CELL/B.E. Architecture.
Implementation of Canny Edge Detector of color images on CELL/B.E. Architecture. Chirag Gupta,Sumod Mohan K cgupta@clemson.edu, sumodm@clemson.edu Abstract In this project we propose a method to improve
More informationAutomatic Caption Localization in Compressed Video
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 22, NO. 4, APRIL 2000 385 Automatic Caption Localization in Compressed Video Yu Zhong, Hongjiang Zhang, and Anil K. Jain, Fellow, IEEE
More informationAutomated System for Computationof Burnt Forest Region using Image Processing
International Journal of Computer Sciences and Engineering Open Access Research Paper Volume-4, Special Issue-1 E-ISSN: 2347-2693 Automated System for Computationof Burnt Forest Region using Image Processing
More informationRobust and accurate global vision system for real time tracking of multiple mobile robots
Robust and accurate global vision system for real time tracking of multiple mobile robots Mišel Brezak Ivan Petrović Edouard Ivanjko Department of Control and Computer Engineering, Faculty of Electrical
More informationFace Model Fitting on Low Resolution Images
Face Model Fitting on Low Resolution Images Xiaoming Liu Peter H. Tu Frederick W. Wheeler Visualization and Computer Vision Lab General Electric Global Research Center Niskayuna, NY, 1239, USA {liux,tu,wheeler}@research.ge.com
More information3D Model based Object Class Detection in An Arbitrary View
3D Model based Object Class Detection in An Arbitrary View Pingkun Yan, Saad M. Khan, Mubarak Shah School of Electrical Engineering and Computer Science University of Central Florida http://www.eecs.ucf.edu/
More informationMonte Carlo Path Tracing
CS294-13: Advanced Computer Graphics Lecture #5 University of California, Berkeley Wednesday, 23 September 29 Monte Carlo Path Tracing Lecture #5: Wednesday, 16 September 29 Lecturer: Ravi Ramamoorthi
More information1 if 1 x 0 1 if 0 x 1
Chapter 3 Continuity In this chapter we begin by defining the fundamental notion of continuity for real valued functions of a single real variable. When trying to decide whether a given function is or
More informationSuper-resolution method based on edge feature for high resolution imaging
Science Journal of Circuits, Systems and Signal Processing 2014; 3(6-1): 24-29 Published online December 26, 2014 (http://www.sciencepublishinggroup.com/j/cssp) doi: 10.11648/j.cssp.s.2014030601.14 ISSN:
More informationIntroduction to Support Vector Machines. Colin Campbell, Bristol University
Introduction to Support Vector Machines Colin Campbell, Bristol University 1 Outline of talk. Part 1. An Introduction to SVMs 1.1. SVMs for binary classification. 1.2. Soft margins and multi-class classification.
More informationModelling 3D Avatar for Virtual Try on
Modelling 3D Avatar for Virtual Try on NADIA MAGNENAT THALMANN DIRECTOR MIRALAB UNIVERSITY OF GENEVA DIRECTOR INSTITUTE FOR MEDIA INNOVATION, NTU, SINGAPORE WWW.MIRALAB.CH/ Creating Digital Humans Vertex
More informationLocating and Decoding EAN-13 Barcodes from Images Captured by Digital Cameras
Locating and Decoding EAN-13 Barcodes from Images Captured by Digital Cameras W3A.5 Douglas Chai and Florian Hock Visual Information Processing Research Group School of Engineering and Mathematics Edith
More informationROBOTRACKER A SYSTEM FOR TRACKING MULTIPLE ROBOTS IN REAL TIME. by Alex Sirota, alex@elbrus.com
ROBOTRACKER A SYSTEM FOR TRACKING MULTIPLE ROBOTS IN REAL TIME by Alex Sirota, alex@elbrus.com Project in intelligent systems Computer Science Department Technion Israel Institute of Technology Under the
More informationApplications to Data Smoothing and Image Processing I
Applications to Data Smoothing and Image Processing I MA 348 Kurt Bryan Signals and Images Let t denote time and consider a signal a(t) on some time interval, say t. We ll assume that the signal a(t) is
More informationAssessment. Presenter: Yupu Zhang, Guoliang Jin, Tuo Wang Computer Vision 2008 Fall
Automatic Photo Quality Assessment Presenter: Yupu Zhang, Guoliang Jin, Tuo Wang Computer Vision 2008 Fall Estimating i the photorealism of images: Distinguishing i i paintings from photographs h Florin
More informationRobert Collins CSE598G. More on Mean-shift. R.Collins, CSE, PSU CSE598G Spring 2006
More on Mean-shift R.Collins, CSE, PSU Spring 2006 Recall: Kernel Density Estimation Given a set of data samples x i ; i=1...n Convolve with a kernel function H to generate a smooth function f(x) Equivalent
More informationWe can display an object on a monitor screen in three different computer-model forms: Wireframe model Surface Model Solid model
CHAPTER 4 CURVES 4.1 Introduction In order to understand the significance of curves, we should look into the types of model representations that are used in geometric modeling. Curves play a very significant
More informationBuilding an Advanced Invariant Real-Time Human Tracking System
UDC 004.41 Building an Advanced Invariant Real-Time Human Tracking System Fayez Idris 1, Mazen Abu_Zaher 2, Rashad J. Rasras 3, and Ibrahiem M. M. El Emary 4 1 School of Informatics and Computing, German-Jordanian
More informationThe Role of Size Normalization on the Recognition Rate of Handwritten Numerals
The Role of Size Normalization on the Recognition Rate of Handwritten Numerals Chun Lei He, Ping Zhang, Jianxiong Dong, Ching Y. Suen, Tien D. Bui Centre for Pattern Recognition and Machine Intelligence,
More informationFriendly Medical Image Sharing Scheme
Journal of Information Hiding and Multimedia Signal Processing 2014 ISSN 2073-4212 Ubiquitous International Volume 5, Number 3, July 2014 Frily Medical Image Sharing Scheme Hao-Kuan Tso Department of Computer
More informationFast Digital Image Inpainting
Appeared in the Proceedings of the International Conference on Visualization, Imaging and Image Processing (VIIP 2001), Marbella, Spain. September 3-5, 2001 Fast Digital Image Inpainting Manuel M. Oliveira
More informationA Novel Hole filling method based on Projection onto Convex Set in DIBR
3rd International Conference on Multimedia Technology ICMT 2013) A Novel Hole filling method based on Projection onto Convex Set in DIBR Weiquan Wang1 and Yingyun Yang2 and Qian Liang3 Abstract. Depth
More informationEnhanced LIC Pencil Filter
Enhanced LIC Pencil Filter Shigefumi Yamamoto, Xiaoyang Mao, Kenji Tanii, Atsumi Imamiya University of Yamanashi {daisy@media.yamanashi.ac.jp, mao@media.yamanashi.ac.jp, imamiya@media.yamanashi.ac.jp}
More informationSkillsUSA 2014 Contest Projects 3-D Visualization and Animation
SkillsUSA Contest Projects 3-D Visualization and Animation Click the Print this Section button above to automatically print the specifications for this contest. Make sure your printer is turned on before
More informationImproving Computer Vision-Based Indoor Wayfinding for Blind Persons with Context Information
Improving Computer Vision-Based Indoor Wayfinding for Blind Persons with Context Information YingLi Tian 1, Chucai Yi 1, and Aries Arditi 2 1 Electrical Engineering Department The City College and Graduate
More informationEdge tracking for motion segmentation and depth ordering
Edge tracking for motion segmentation and depth ordering P. Smith, T. Drummond and R. Cipolla Department of Engineering University of Cambridge Cambridge CB2 1PZ,UK {pas1001 twd20 cipolla}@eng.cam.ac.uk
More informationMachine vision systems - 2
Machine vision systems Problem definition Image acquisition Image segmentation Connected component analysis Machine vision systems - 1 Problem definition Design a vision system to see a flat world Page
More informationDegree Reduction of Interval SB Curves
International Journal of Video&Image Processing and Network Security IJVIPNS-IJENS Vol:13 No:04 1 Degree Reduction of Interval SB Curves O. Ismail, Senior Member, IEEE Abstract Ball basis was introduced
More informationReal-time Traffic Congestion Detection Based on Video Analysis
Journal of Information & Computational Science 9: 10 (2012) 2907 2914 Available at http://www.joics.com Real-time Traffic Congestion Detection Based on Video Analysis Shan Hu a,, Jiansheng Wu a, Ling Xu
More informationDual Methods for Total Variation-Based Image Restoration
Dual Methods for Total Variation-Based Image Restoration Jamylle Carter Institute for Mathematics and its Applications University of Minnesota, Twin Cities Ph.D. (Mathematics), UCLA, 2001 Advisor: Tony
More informationHow To Segmentate An Image
Edge Strength Functions as Shape Priors in Image Segmentation Erkut Erdem, Aykut Erdem, and Sibel Tari Middle East Technical University, Department of Computer Engineering, Ankara, TR-06531, TURKEY, {erkut,aykut}@ceng.metu.edu.tr,
More informationDYNAMIC DOMAIN CLASSIFICATION FOR FRACTAL IMAGE COMPRESSION
DYNAMIC DOMAIN CLASSIFICATION FOR FRACTAL IMAGE COMPRESSION K. Revathy 1 & M. Jayamohan 2 Department of Computer Science, University of Kerala, Thiruvananthapuram, Kerala, India 1 revathysrp@gmail.com
More informationAutomatic Traffic Estimation Using Image Processing
Automatic Traffic Estimation Using Image Processing Pejman Niksaz Science &Research Branch, Azad University of Yazd, Iran Pezhman_1366@yahoo.com Abstract As we know the population of city and number of
More informationMorphological segmentation of histology cell images
Morphological segmentation of histology cell images A.Nedzved, S.Ablameyko, I.Pitas Institute of Engineering Cybernetics of the National Academy of Sciences Surganova, 6, 00 Minsk, Belarus E-mail abl@newman.bas-net.by
More informationComputational Geometry Lab: FEM BASIS FUNCTIONS FOR A TETRAHEDRON
Computational Geometry Lab: FEM BASIS FUNCTIONS FOR A TETRAHEDRON John Burkardt Information Technology Department Virginia Tech http://people.sc.fsu.edu/ jburkardt/presentations/cg lab fem basis tetrahedron.pdf
More informationOverview. Essential Questions. Precalculus, Quarter 4, Unit 4.5 Build Arithmetic and Geometric Sequences and Series
Sequences and Series Overview Number of instruction days: 4 6 (1 day = 53 minutes) Content to Be Learned Write arithmetic and geometric sequences both recursively and with an explicit formula, use them
More informationCUBE-MAP DATA STRUCTURE FOR INTERACTIVE GLOBAL ILLUMINATION COMPUTATION IN DYNAMIC DIFFUSE ENVIRONMENTS
ICCVG 2002 Zakopane, 25-29 Sept. 2002 Rafal Mantiuk (1,2), Sumanta Pattanaik (1), Karol Myszkowski (3) (1) University of Central Florida, USA, (2) Technical University of Szczecin, Poland, (3) Max- Planck-Institut
More informationExtracting a Good Quality Frontal Face Images from Low Resolution Video Sequences
Extracting a Good Quality Frontal Face Images from Low Resolution Video Sequences Pritam P. Patil 1, Prof. M.V. Phatak 2 1 ME.Comp, 2 Asst.Professor, MIT, Pune Abstract The face is one of the important
More informationFCE: A Fast Content Expression for Server-based Computing
FCE: A Fast Content Expression for Server-based Computing Qiao Li Mentor Graphics Corporation 11 Ridder Park Drive San Jose, CA 95131, U.S.A. Email: qiao li@mentor.com Fei Li Department of Computer Science
More informationMVA ENS Cachan. Lecture 2: Logistic regression & intro to MIL Iasonas Kokkinos Iasonas.kokkinos@ecp.fr
Machine Learning for Computer Vision 1 MVA ENS Cachan Lecture 2: Logistic regression & intro to MIL Iasonas Kokkinos Iasonas.kokkinos@ecp.fr Department of Applied Mathematics Ecole Centrale Paris Galen
More informationAutomatic Labeling of Lane Markings for Autonomous Vehicles
Automatic Labeling of Lane Markings for Autonomous Vehicles Jeffrey Kiske Stanford University 450 Serra Mall, Stanford, CA 94305 jkiske@stanford.edu 1. Introduction As autonomous vehicles become more popular,
More informationModel Repair. Leif Kobbelt RWTH Aachen University )NPUT $ATA 2EMOVAL OF TOPOLOGICAL AND GEOMETRICAL ERRORS !NALYSIS OF SURFACE QUALITY
)NPUT $ATA 2ANGE 3CAN #!$ 4OMOGRAPHY 2EMOVAL OF TOPOLOGICAL AND GEOMETRICAL ERRORS!NALYSIS OF SURFACE QUALITY 3URFACE SMOOTHING FOR NOISE REMOVAL 0ARAMETERIZATION 3IMPLIFICATION FOR COMPLEXITY REDUCTION
More informationImage Normalization for Illumination Compensation in Facial Images
Image Normalization for Illumination Compensation in Facial Images by Martin D. Levine, Maulin R. Gandhi, Jisnu Bhattacharyya Department of Electrical & Computer Engineering & Center for Intelligent Machines
More informationTracking Moving Objects In Video Sequences Yiwei Wang, Robert E. Van Dyck, and John F. Doherty Department of Electrical Engineering The Pennsylvania State University University Park, PA16802 Abstract{Object
More informationSo which is the best?
Manifold Learning Techniques: So which is the best? Todd Wittman Math 8600: Geometric Data Analysis Instructor: Gilad Lerman Spring 2005 Note: This presentation does not contain information on LTSA, which
More informationTransparency and Occlusion
Transparency and Occlusion Barton L. Anderson University of New South Wales One of the great computational challenges in recovering scene structure from images arises from the fact that some surfaces in
More informationHow To Fix Out Of Focus And Blur Images With A Dynamic Template Matching Algorithm
IJSTE - International Journal of Science Technology & Engineering Volume 1 Issue 10 April 2015 ISSN (online): 2349-784X Image Estimation Algorithm for Out of Focus and Blur Images to Retrieve the Barcode
More information3. Interpolation. Closing the Gaps of Discretization... Beyond Polynomials
3. Interpolation Closing the Gaps of Discretization... Beyond Polynomials Closing the Gaps of Discretization... Beyond Polynomials, December 19, 2012 1 3.3. Polynomial Splines Idea of Polynomial Splines
More information521466S Machine Vision Assignment #7 Hough transform
521466S Machine Vision Assignment #7 Hough transform Spring 2014 In this assignment we use the hough transform to extract lines from images. We use the standard (r, θ) parametrization of lines, lter the
More informationOBJECT TRACKING USING LOG-POLAR TRANSFORMATION
OBJECT TRACKING USING LOG-POLAR TRANSFORMATION A Thesis Submitted to the Gradual Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements
More informationLevel Set Evolution Without Re-initialization: A New Variational Formulation
Level Set Evolution Without Re-initialization: A New Variational Formulation Chunming Li 1, Chenyang Xu 2, Changfeng Gui 3, and Martin D. Fox 1 1 Department of Electrical and 2 Department of Imaging 3
More information