Image Processing IEEE 2014 Projects

Web : www.kasanpro.com     Email : sales@kasanpro.com

List Link : http://kasanpro.com/projects-list/image-processing-ieee-2014-projects

Title :Building Change Detection Based on Satellite Stereo Imagery and Digital Surface ModelsLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/building-change-detection-based-satellite-stereo-imagery-digital-surface-model

Abstract : Building change detection is a major issue for urban area monitoring. Due to different imaging conditionsand sensor parameters, 2-D information delivered by satellite images from different dates is often not sufficient whendealing with building changes. Moreover, due to the similar spectral characteristics, it is often difficult to distinguishbuildings from other man-made constructions, like roads and bridges, during the change detection procedure.Therefore, stereo imagery is of importance to provide the height component which is very helpful in analyzing 3-Dbuilding changes. In this paper, we propose a change detection method based on stereo imagery and digital surfacemodels (DSMs) generated with stereo matching methodology and provide a solution by the joint use of heightchanges and Kullback-Leibler divergence similarity measure between the original images. The Dempster-Shaferfusion theory is adopted to combine these two change indicators to improve the accuracy. In addition, vegetation andshadow classifications are used as no-building change indicators for refining the change detection results. In the end,an object-based building extraction method based on shape features is performed. For evaluation purpose, theproposed method is applied in two test areas, one is in an industrial area in Korea with stereo imagery from the samesensor and the other represents a dense urban area in Germany using stereo imagery from different sensors withdifferent resolutions. Our experimental results confirm the efficiency and high accuracy of the proposed methodologyeven for different kinds and combinations of stereo images and consequently different DSM qualities.

Title :PCA Feature Extraction for Change Detection in Multidimensional Unlabelled DataLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/pca-feature-extraction-change-detection-multidimensional-unlabelled-data

Abstract : When classifiers are deployed in real world applications, it is assumed that the distribution of the incomingdata matches the distribution of the data used to train the classifier. This assumption is often incorrect, whichnecessitates some form of change detection or adaptive classification. While there is a lot of research on changedetection based on the classification error, monitored over the course of the operation of the classifier, findingchanges in multidimensional unlabelled data is still a challenge. Here we propose to apply principal componentanalysis (PCA) for feature extraction prior to the change detection. Supported by a theoretical example, we argue thatthe components with the lowest variance should be retained as the extracted features because they are more likely tobe affected by a change. We chose a recently proposed semi-parametric log-likelihood change detection criterion(SPLL) which is sensitive to changes in both mean and variance of the multidimensional distribution. An experimentwith 35 data sets and an illustration with a simple video segmentation demonstrate the advantage of using extractedfeatures compared to raw data. Further analysis shows that feature extraction through PCA is beneficial, specificallyfor data with multiple balanced classes.

Title :Remote Sensing Image Segmentation by Combining Spectral and Texture FeaturesLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/remote-sensing-image-segmentation-combining-spectral-texture-features

Abstract : We present a new method for remote sensing image segmentation, which utilizes both spectral andtexture information. Linear filters are used to provide enhanced spatial patterns. For each pixel location, we computecombined spectral and texture features using local spectral histograms, which concatenate local histograms of allinput bands. We regard each feature as a linear combination of several representative features, each of whichcorresponds to a segment. Segmentation is given by estimating combination weights, which indicate segmentownership of pixels. We present segmentation solutions where representative features are either known or unknown.

We also show that feature dimensions can be greatly reduced via subspace projection. The scale issue isinvestigated, and an algorithm is presented to automatically select proper scales, which does not requiresegmentation at multiplescale levels. Experimental results demonstrate the promise of the proposed method.

Title :Brain Tumor Detection using Region based Iterative Reconstruction and SegmentationLanguage : Java

Project Link : http://kasanpro.com/p/java/brain-tumor-detection-region-based-iterative-reconstruction-segmentation

Abstract : X-ray computed tomography (CT) is a powerful tool for noninvasive imaging of time-varying objects.Identifiying tumors from the CT image is a chalanging one. In this paper we proposed a reconstruct method for CTimage and tumors are detected then using edge based segmentation algorithm.. In the past, methods have beenproposed to reconstruct images from continuously changing objects. For discretely or structurally changing objects,however, such methods fail to reconstruct high quality images, mainly because assumptions about continuity are nolonger valid. In this paper, we propose a method to reconstruct structurally changing objects. Starting from theobservation that there exist regions within the scanned object that remain unchanged over time, we introduce aniterative optimization routine that can automatically determine these regions and incorporate this knowledge in analgebraic reconstruction method. And tumor detection was made from the reconstructed image.

Title :Automatic graph based approach for prior detection of diabetes and hypertension in retinal imagesLanguage : Java

Project Link : http://kasanpro.com/p/java/automatic-graph-based-prior-detection-diabetes-hypertension-retinal-images

Abstract : Retinal vessels are affected by several systemic diseases, namely diabetes, hypertension, and vasculardisorders. In diabetic retinopathy, the blood vessels often show abnormalities at early stages, as well as vesseldiameter alterations . Changes in retinal blood vessels, such as significant dilatation and elongation of main arteries,veins, and their branches are also frequently associated with hypertension and other cardiovascular pathologies. Theclassification of retinal vessels into artery/vein (A/V) is an important phase for automating the detection of vascularchanges, and for the calculation of characteristic signs associated with several systemic diseases such as diabetes,hypertension, and other cardiovascular conditions. This paper presents an automatic approach for A/V classificationbased on the analysis of a graph extracted from the retinal vasculature. The proposed method classifies the entirevascular tree deciding on the type of each intersection point (graph nodes) and assigning one of two labels to eachvessel segment (graph links). Final classification of a vessel segment as A/V is performed through the combination ofthe graph-based labeling results with a set of intensity features. The features were extracted, including exudates,bifurcation angle, artery-to-veins diameter ratio, mean artery and veins diameters, form and size of optic disc, andvessel tortuosity. And the identification of diabetes are made by the rule based conditions.

Image Processing IEEE 2014 Projects

Title :A Compressive Sensing based Secure Watermark Detection and Privacy Preserving Storage FrameworkLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/secure-watermark-detection-privacy-preserving-storage-framework

Abstract : Privacy is a critical issue when the data owners outsource data storage or processing to a third partycomputing service, such as the cloud. In this paper, we identify a cloud computing application scenario that requiressimultaneously performing secure watermark detection and privacy preserving multimedia data storage. We thenpropose a compressive sensing (CS)-based framework using secure multiparty computation (MPC) protocols toaddress such a requirement. In our framework, the multimedia data and secret watermark pattern are presented tothe cloud for secure watermark detection in a CS domain to protect the privacy. During CS transformation, the privacyof the CS matrix and the watermark pattern is protected by the MPC protocols under the semi-honest security model.We derive the expected watermark detection performance in the CS domain, given the target image, watermarkpattern, and the size of the CS matrix (but without the CS matrix itself). The correctness of the derived performancehas been validated by our experiments. Our theoretical analysis and experimental results show that secure watermarkdetection in the CS domain is feasible. Our framework can also be extended to other collaborative secure signalprocessing and data-mining applications in the cloud.

Title :A New Iterative Triclass Thresholding Technique in Image SegmentationLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/new-iterative-triclass-thresholding-technique-image-segmentation

Abstract : We present a new method in image segmentation that is based on Otsu's method but iteratively searchesfor subregions of the image for segmentation, instead of treating the full image as a whole region for processing. Theiterative method starts with Otsu's threshold and computes the mean values of the two classes as separated by thethreshold. Based on the Otsu's threshold and the two mean values, the method separates the image into threeclasses instead of two as the standard Otsu's method does. The first two classes are determined as the foregroundand background and they will not be processed further. The third class is denoted as a to-be-determined (TBD) regionthat is processed at next iteration. At the succeeding iteration, Otsu's method is applied on the TBD region tocalculate a new threshold and two class means and the TBD region is again separated into three classes, namely,foreground, background, and a new TBD region, which by definition is smaller than the previous TBD regions. Then,the new TBD region is processed in the similar manner. The process stops when the Otsu's thresholds calculatedbetween two iterations is less than a preset threshold. Then, all the intermediate foreground and background regionsare, respectively, combined to create the final segmentation result. Tests on synthetic and real images showed thatthe new iterative method can achieve better performance than the standard Otsu's method in many challengingcases, such as identifying weak objects and revealing fine structures of complex objects while the addedcomputational cost is minimal.

Title :As-Projective-As-Possible Image Stitching with Moving DLTLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/as-projective-as-possible-image-stitching-moving-dlt

Abstract : We investigate projective estimation under model inadequacies, i.e., when the underpinning assumptionsof the projective model are not fully satisfied by the data. We focus on the task of image stitching which is customarilysolved by estimating a projective warp -- a model that is justified when the scene is planar or when the views differpurely by rotation. Such conditions are easily violated in practice, and this yields stitching results with ghostingartefacts that necessitate the usage of deghosting algorithms. To this end we propose as-projective-as-possiblewarps, i.e., warps that aim to be globally projective, yet allow local non-projective deviations to account for violationsto the assumed imaging conditions. Based on a novel estimation technique called Moving Direct LinearTransformation (Moving DLT), our method seamlessly bridges image regions that are inconsistent with the projectivemodel. The result is highly accurate image stitching, with significantly reduced ghosting effects, thus lowering thedependency on post hoc deghosting.

http://kasanpro.com/ieee/final-year-project-center-kanniyakumari-reviews

Title :Captcha as Graphical Passwords--A New Security Primitive Based on Hard AI ProblemsLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/captcha-graphical-password

Abstract : Many security primitives are based on hard mathematical problems. Using hard AI problems for security isemerging as an exciting new paradigm, but has been underexplored. In this paper, we present a new securityprimitive based on hard AI problems, namely, a novel family of graphical password systems built on top of Captchatechnology, which we call Captcha as graphical passwords (CaRP). CaRP is both a Captcha and a graphicalpassword scheme. CaRP addresses a number of security problems altogether, such as online guessing attacks, relayattacks, and, if combined with dual-view technologies, shoulder-surfing attacks. Notably, a CaRP password can befound only probabilistically by automatic online guessing attacks even if the password is in the search set. CaRP alsooffers a novel approach to address the well-known image hotspot problem in popular graphical password systems,such as PassPoints, that often leads to weak password choices. CaRP is not a panacea, but it offers reasonablesecurity and usability and appears to fit well with some practical applications for improving online security.

Title :Corruptive Artifacts Suppression for Example-Based Color TransferLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/corruptive-artifacts-suppression-example-based-color-transfer

Abstract : Example-based color transfer is a critical operation in image editing but easily suffers from some corruptiveartifacts in themapping process. In this paper,we propose a novel unified color transfer framework with corruptiveartifacts suppression, which performs iterative probabilistic color mapping with self-learning filtering scheme andmultiscale detail manipulation scheme inminimizing the normalized Kullback-Leibler distance. First, an iterativeprobabilistic color mapping is applied to construct the mapping relationship between the reference and target images.Then, a self-learning filtering scheme is applied into the transfer process to prevent from artifacts and extract details.The transferred output and the extracted multi-levels details are integrated by the measurement minimization to yield

the final result. Our framework achieves a sound grain suppression, color fidelity and detail appearance seamlessly.For demonstration, a series of objective and subjective measurements are used to evaluate the quality in colortransfer. Finally, a few extended applications are implemented to show the applicability of this framework.

Image Processing IEEE 2014 Projects

Title :Fingerprint Compression Based on Sparse RepresentationLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/fingerprint-compression-based-sparse-representation

Abstract : A new fingerprint compression algorithm based on sparse representation is introduced. Obtaining anovercomplete dictionary from a set of fingerprint patches allows us to represent them as a sparse linear combinationof dictionary atoms. In the algorithm, we first construct a dictionary for predefined fingerprint image patches. For anew given fingerprint images, represent its patches according to the dictionary by computing l0-minimization and thenquantize and encode the representation. In this paper, we consider the effect of various factors on compressionresults. Three groups of fingerprint images are tested. The experiments demonstrate that our algorithm is efficientcompared with several competing compression techniques (JPEG, JPEG 2000, andWSQ), especially at highcompression ratios. The experiments also illustrate that the proposed algorithm is robust to extract minutiae.

Title :How to Estimate the Regularization Parameter for Spectral Regression Discriminant Analysis and its KernelVersion?Language : Matlab

Project Link : http://kasanpro.com/p/matlab/regularization-parameter-spectral-regression-discriminant-analysis

Abstract : Spectral regression discriminant analysis (SRDA) has recently been proposed as an efficient solution tolarge-scale subspace learning problems. There is a tunable regularization parameter in SRDA, which is critical toalgorithm performance. However, how to automatically set this parameter has not been well solved until now. So thisregularization parameter was only set to be a constant in SRDA, which is obviously suboptimal. This paper proposesto automatically estimate the optimal regularization parameter of SRDA based on the perturbation linear discriminantanalysis (PLDA). In addition, two parameter estimation methods for the kernel version of SRDA are also developed.One is derived from the method of optimal regularization parameter estimation for SRDA. The other is to utilize thekernel version of PLDA. Experiments on a number of publicly available databases demonstrate the effectiveness ofthe proposed methods for face recognition, spoken letter recognition, handwritten digit recognition, and textcategorization.

Title :Image Classification Using Multiscale Information Fusion Based on Saliency Driven Nonlinear Diffusion FilteringLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/image-classification-based-fusion-diffusion-filtering

Abstract : In this paper, we propose saliency driven image multiscale nonlinear diffusion filtering. The resulting scalespace in general preserves or even enhances semantically important structures such as edges, lines, or flow-likestructures in the foreground, and inhibits and smoothes clutter in the background. The image is classified usingmultiscale information fusion based on the original image, the image at the final scale at which the diffusion processconverges, and the image at a midscale. Our algorithm emphasizes the foreground features, which are important forimage classification. The background image regions, whether considered as contexts of the foreground or noise to theforeground, can be globally handled by fusing information from different scales. Experimental tests of theeffectiveness of the multiscale space for the image classification are conducted on the following publicly availabledatasets: 1) the PASCAL 2005 dataset; 2) the Oxford 102 flowers dataset; and 3) the Oxford 17 flowers dataset, withhigh classification rates.

Title :LBP-Based Edge-Texture Features for Object RecognitionLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/lbp-based-edge-texture-features-object-recognition

Abstract : This paper proposes two sets of novel edge-texture features, Discriminative Robust Local Binary Pattern(DRLBP) and Ternary Pattern (DRLTP), for object recognition. By investigating the limitations of Local Binary Pattern(LBP), Local Ternary Pattern (LTP) and Robust LBP (RLBP), DRLBP and DRLTP are proposed as new features.They solve the problem of discrimination between a bright object against a dark background and vice-versa inherent

in LBP and LTP. DRLBP also resolves the problem of RLBP whereby LBP codes and their complements in the sameblock are mapped to the same code. Furthermore, the proposed features retain contrast information necessary forproper representation of object contours that LBP, LTP, and RLBP discard. Our proposed features are tested onseven challenging data sets: INRIA Human, Caltech Pedestrian, UIUC Car, Caltech 101, Caltech 256, Brodatz, andKTH-TIPS2- a. Results demonstrate that the proposed features outperform the compared approaches on most datasets.

Title :Learning Layouts for Single-Page Graphic DesignsLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/learning-layouts-single-page-graphic-designs

Abstract : This paper presents an approach for automatically creating graphic design layouts using a newenergy-based model derived from design principles. The model includes several new algorithm for analyzing graphicdesigns, including the prediction of perceived importance, alignment detection, and hierarchical segmentation. Giventhe model, we use optimization to synthesize new layouts for a variety of single-page graphic designs. Modelparameters are learned with Nonlinear Inverse Optimization (NIO) from a small number of example layouts. Todemonstrate our approach, we show result for application including generation design layouts in various styles,retargeting designs to new sizes, and improving existing designs. We also compare our automatic results withdesigns created using crowdsourcing and show that our approach performs slightly better than novice designers.

Image Processing IEEE 2014 Projects

Title :Localization of License Plate Number Using Dynamic Image Processing Techniques And Genetic AlgorithmsLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/license-plate-number-localization-using-image-processing-and-genetic-algorithms

Abstract : In this research, a design of a new genetic algorithm (GA) is introduced to detect the locations of theLicense Plate (LP) symbols. An adaptive threshold method has been applied to overcome the dynamic changes ofillumination conditions when converting the image into binary. Connected component analysis technique (CCAT) isused to detect candidate objects inside the unknown image. A scale-invariant Geometric Relationship Matrix (GRM)has been introduced to model the symbols layout in any LP which simplifies system adaptability when applied indifferent countries. Moreover, two new crossover operators, based on sorting, have been introduced which greatlyimproved the convergence speed of the system. Most of CCAT problems such as touching or broken bodies havebeen minimized by modifying the GA to perform partial match until reaching to an acceptable fitness value. Thesystem has been implemented using MATLAB and various image samples have been experimented to verify thedistinction of the proposed system. Encouraging results with 98.4% overall accuracy have been reported for twodifferent datasets having variability in orientation, scaling, plate location, illumination and complex background.Examples of distorted plate images were successfully detected due to the independency on the shape, color, orlocation of the plate.

http://kasanpro.com/ieee/final-year-project-center-kanniyakumari-reviews

Title :Model-Based Edge Detector for Spectral Imagery Using Sparse Spatiospectral MasksLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/model-based-edge-detector-spectral-imagery-using-sparse-spatiospectral-masks

Abstract : Two model-based algorithms for edge detection in spectral imagery are developed that specifically targetcapturing intrinsic features such as isoluminant edges that are characterized by a jump in color but not in intensity.Given prior knowledge of the classes of reflectance or emittance spectra associated with candidate objects in ascene, a small set of spectral-band ratios, which most profoundly identify the edge between each pair of materials,are selected to define a edge signature. The bands that form the edge signature are fed into a spatial mask,producing a sparse joint spatiospectral nonlinear operator. The first algorithm achieves edge detection for everymaterial pair by matching the response of the operator at every pixel with the edge signature for the pair of materials.The second algorithm is a classifier-enhanced extension of the first algorithm that adaptively accentuates distinctivefeatures before applying the spatiospectral operator. Both algorithms are extensively verified using spectral imageryfrom the airborne hyperspectral imager and from a dots-in-a-well midinfrared imager. In both cases, the multicolorgradient (MCG) and the hyperspectral/spatial detection of edges (HySPADE) edge detectors are used as a

benchmark for comparison. The results demonstrate that the proposed algorithms outperform the MCG andHySPADE edge detectors in accuracy, especially when isoluminant edges are present. By requiring only a few bandsas input to the spatiospectral operator, the algorithms enable significant levels of data compression in band selection.In the presented examples, the required operations per pixel are reduced by a factor of 71 with respect to thoserequired by the MCG edge detector.

Title :Modeling of Speaking Rate Influences on Mandarin Speech Prosody and Its Application to SpeakingRate-controlled TTSLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/application-speaking-rate-controlled-tts

Abstract : A new data-driven approach to building a speaking rate-dependent hierarchical prosodic model (SR-HPM),directly from a large prosody-unlabeled speech database containing utterances of various speaking rates, to describethe influences of speaking rate on Mandarin speech prosody is proposed. It is an extended version of the existingHPM model which contains 12 sub-models to describe various relationships of prosodic-acoustic features of speechsignal, linguistic features of the associated text, and prosodic tags representing the prosodic structure of speech. Twomain modifications are suggested. One is designing proper normalization functions from the statistics of the wholedatabase to compensate the influences of speaking rate on all prosodic-acoustic features. Another is modifying theHPM training to let its parameters be speaking-rate dependent. Experimental results on a large Mandarin read speechcorpus showed that the parameters of the SR-HPM together with these feature normalization functions interpreted theeffects of speaking rate onMandarin speech prosody very well. An application of the SR-HPM to design andimplement a speaking rate-controlledMandarin TTS system is demonstrated. The system can generate naturalsynthetic speech for any given speaking rate in awide range of 3.4-6.8 syllables/sec.Two subjective tests,MOSandpreference test,were conducted to compare the proposed system with the popular HTS system. TheMOS scores ofthe proposed system were in the range of 3.58-3.83 for eight different speaking rates, while they were in 3.09-3.43 forHTS. Besides, the proposed system had higher preference scores (49.8%-79.6%) than those (9.8%-30.7%) ofHTS.This confirmed the effectiveness of the speaking rate control method of the proposed TTS system.

Title :Multisensor Fusion-Based Concurrent Environment Mapping and Moving Object Detection for IntelligentService RoboticsLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/moving-object-detection-intelligent-service-robotics

Abstract : Intelligent service robot development is an important and critical issue for human community applications.With the diverse and complex service needs, the perception and navigation are essential subjects. This investigationfocuses on the synergistic fusion of multiple sensors for an intelligent service robot that not only performsself-localization and mapping but also detects moving objects or people in the building it services. First of all, a newaugmented approach of graph-based optimal estimation was derived for concurrent robot postures and moving objecttrajectory estimate. Moreover, all the moving object detection issues of a robot's indoor navigation are divided andconquered via multisensor fusion methodologies. From bottom to up, the estimation fusion methods are tacticallyutilized to get a more precise result than the one from only the laser ranger or stereo vision. Furthermore, for solvingthe consistent association problem of moving objects, a covariance area intersection belief assignment is applied formotion state evaluation and the complementary evidences such as kinematics and vision features are bothsynergized together to enhance the association efficiency with the evidence fusion method. The proof of concept withexperiments has been successfully demonstrated and analyzed.

Title :Personalized Geo-Specific Tag Recommendation for Photos on Social WebsitesLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/personalized-geo-specific-tag-recommendation-photos-social-websites

Abstract : Social tagging becomes increasingly important to organize and search large-scale community-contributedphotos on social websites. To facilitate generating high-quality social tags, tag recommendation by automaticallyassigning relevant tags to photos draws particular research interest. In this paper, we focus on the personalized tagrecommendation task and try to identify user-preferred, geo-location-specific as well as semantically relevant tags fora photo by leveraging rich contexts of the freely available community-contributed photos. For users and geo-locations,we assume they have different preferred tags assigned to a photo, and propose a subspace learning method toindividually uncover the both types of preferences. The goal of our work is to learn a unified subspace shared by thevisual and textual domains to make visual features and textual information of photos comparable. Considering thevisual feature is a lower level representation on semantics than the textual information, we adopt a progressivelearning strategy by additionally introducing an intermediate subspace for the visual domain, and expect it to haveconsistent local structure with the textual space. Accordingly, the unified subspace is mapped from the intermediatesubspace and the textual space respectively. We formulate the above learning problems into a united form, and

present an iterative optimization with its convergence proof. Given an untagged photo with its geo-location to a user,the user-preferred and the geo-location-specific tags are found by the nearest neighbor search in the correspondingunified spaces. Then we combine the obtained tags and the visual appearance of the photo to discover thesemantically and visually related photos, among which the most frequent tags are used as the recommended tags.Experiments on a large-scale data set collected from Flickr verify the effectivity of the proposed solution.

Image Processing IEEE 2014 Projects

Title :Photometric Stereo Using Sparse Bayesian Regression for General Diffuse SurfacesLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/photometric-stereo-using-sparse-bayesian-regression-general-diffuse-surfaces

Abstract : Most conventional algorithms for non-Lambertian photometric stereo can be partitioned into twocategories. The first category is built upon stable outlier rejection techniques while assuming a dense Lambertianstructure for the inliers, and thus performance degrades when general diffuse regions are present. The second utilizescomplex reflectance representations and non-linear optimization over pixels to handle non-Lambertian surfaces, butdoes not explicitly account for shadows or other forms of corrupting outliers. In this paper, we present a purelypixel-wise photometric stereo method that stably and efficiently handles various non-Lambertian effects by assumingthat appearances can be decomposed into a sparse, non-diffuse component (e.g., shadows, specularities, etc.) and adiffuse component represented by a monotonic function of the surface normal and lighting dot-product. This functionis constructed using a piecewise linear approximation to the inverse diffuse model, leading to closed-form estimatesof the surface normals and model parameters in the absence of non-diffuse corruptions. The latter are modeled aslatent variables embedded within a hierarchical Bayesian model such that we may accurately compute the unknownsurface normals while simultaneously separating diffuse from non-diffuse components. Extensive evaluations areperformed that show state-of-the-art performance using both synthetic and real-world images.

Title :Quality Assessment of Stereoscopic 3D Image Compression by Binocular Integration BehaviorsLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/quality-assessment-stereoscopic-3d-image-compression-binocular-integration-behaviors

Abstract : The objective approaches of 3D image quality assessment play a key role for the development ofcompression standards and various 3D multimedia applications. The quality assessment of 3D images faces morenew challenges, such as asymmetric stereo compression, depth perception, and virtual view synthesis, than its 2Dcounterparts. In addition, the widely used 2D image quality metrics (e.g., PSNR and SSIM) cannot be directly appliedto deal with these newly introduced challenges. This statement can be verified by the low correlation between thecomputed objective measures and the subjectively measured mean opinion scores (MOSs), when 3D images are thetested targets. In order to meet these newly introduced challenges, in this paper, besides traditional 2D imagemetrics, the binocular integration behaviors--the binocular combination and the binocular frequency integration, areutilized as the bases for measuring the quality of stereoscopic 3D images. The effectiveness of the proposed metricsis verified by conducting subjective evaluations on publicly available stereoscopic image databases. Experimentalresults show that significant consistency could be reached between the measured MOS and the proposed metrics, inwhich the correlation coefficient between them can go up to 0.88. Furthermore, we found that the proposed metricscan also address the quality assessment of the synthesized color-plusdepth 3D images well. Therefore, it is our beliefthat the binocular integration behaviors are important factors in the development of objective quality assessment for3D images.

Title :Robust Semi-Automatic Depth Map Generation in Unconstrained Images and Video Sequences for 2D toStereoscopic 3D ConversionLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/depth-map-generation-unconstrained-images

Abstract : We describe a system for robustly estimating synthetic depth maps in unconstrained images and videos,for semi-automatic conversion into stereoscopic 3D. Currently, this process is automatic or done manually byrotoscopers. Automatic is the least labor intensive, but makes user intervention or error correction difficult. Manual isthe most accurate, but time consuming and costly. Noting the merits of both, a semi-automatic method blends themtogether, allowing for faster and accurate conversion. This requires user-defined strokes on the image, or over severalkeyframes for video, corresponding to a rough estimate of the depths. After, the rest of the depths are determined,creating depth maps to generate stereoscopic 3D content, with Depth Image Based Rendering to generate theartificial views. Depth map estimation can be considered as a multi-label segmentation problem: each class is adepth. For video, we allow the user to label only the first frame, and we propagate the strokes using computer vision

techniques. We combine the merits of two well-respected segmentation algorithms: Graph Cuts and Random Walks.The diffusion from Random Walks, with the edge preserving of Graph Cuts should give good results. We generategood quality content, more suitable for perception, compared to a similar framework.

Title :Sharing Visual Secrets in Single Image Random Dot StereogramsLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/sharing-visual-secrets-single-image-random-dot-stereograms

Abstract : Visual cryptography schemes (VCSs) generate random and meaningless shares to share and protectsecret images. Conventional VCSs suffer from a transmission risk problem because the noise-like shares will raisethe suspicion of attackers and the attackers might intercept the transmission. Previous research has involved in hidingshared content in halftone shares to reduce these risks, but this method exacerbates the pixel expansion problem andvisual quality degradation problem for recovered images. In this paper, a binocular VCS (BVCS), called the (2,n)-BVCS, and an encryption algorithm are proposed to hide the shared pixels in the single image random dotstereograms (SIRDSs). Because the SIRDSs have the same 2D appearance as the conventional shares of a VCS,this paper tries to use SIRDSs as cover images of the shares of VCSs to reduce the transmission risk of the shares.The encryption algorithm alters the random dots in the SIRDSs according to the construction rule of the (2, n)-BVCSto produce nonpixelexpansion shares of the BVCS. Altering the dots in a SIRDS will degrade the visual quality of thereconstructed 3D objects. Hence, we propose an optimization model that is based on the visual quality requirement ofSIRDSs to develop construction rules for a (2, n)-BVCS that maximize the contrast of the recovered image in theBVCS.

http://kasanpro.com/ieee/final-year-project-center-kanniyakumari-reviews

Title :Single-Image Superresolution of Natural Stochastic Textures Based on Fractional Brownian MotionLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/superresolution-natural-stochastic-textures-based-fractional-brownian-motion

Abstract : Texture enhancement presents an ongoing challenge, in spite of the considerable progress made in recentyears. Whereas most of the effort has been devoted so far to enhancement of regular textures, stochastic texturesthat are encountered in most natural images, still pose an outstanding problem. The purpose of enhancement ofstochastic textures is to recover details, which were lost during the acquisition of the image. In this paper, a texturemodel, based on fractional Brownian motion (fBm), is proposed. The model is global and does not entail using imagepatches. The fBm is a self-similar stochastic process. Self-similarity is known to characterize a large class of naturaltextures. The fBm-based model is evaluated and a single-image regularized superresolution algorithm is derived. Theproposed algorithm is useful for enhancement of a wide range of textures. Its performance is compared withsingle-image superresolution methods and its advantages are highlighted.

Image Processing IEEE 2014 Projects

Title :Weighted KPCA Degree of Homogeneity Amended Nonclassical Receptive Field Inhibition Model for SalientContour Extraction in Low-Light-Level ImageLanguage : Matlab

Project Link : http://kasanpro.com/p/matlab/nonclassical-receptive-field-inhibition-model-salient-contour-extraction

Abstract : The stimulus response of the classical receptive field (CRF) of neuron in primary visual cortex is affectedby its periphery [i.e., non-CRF (nCRF)]. This modulation exerts inhibition, which depends primarily on the correlationof both visual stimulations. The theory of periphery and center interaction with visual characteristics can be applied innight vision information processing. In this paper, a weighted kernel principal component analysis (WKPCA) degree ofhomogeneity (DH) amended inhibition model inspired by visual perceptual mechanisms is proposed to extract salientcontour from complex natural scene in low-light-level image. The core idea is that multifeature analysis can recognizethe homogeneity in modulation coverage effectively. Computationally, a novel WKPCA algorithm is presented toeliminate outliers and anomalous distribution in CRF and accomplish principal component analysis precisely. On thisbasis, a new concept and computational procedure for DH is defined to evaluate the dissimilarity between peripheryand center comprehensively. Through amending the inhibition from nCRF to CRF by DH, our model can reduce theinterference of noises, suppress details, and textures in homogeneous regions accurately. It helps to further avoidmutual suppression among inhomogeneous regions and contour elements. This paper provides an improvedcomputational visual model with high-performance for contour detection from cluttered natural scene in night vision

image.

Title :Region-Based Iterative Reconstruction of Structurally Changing Objects in CTLanguage : Java

Project Link : http://kasanpro.com/p/java/region-based-iterative-reconstruction-structurally-changing-objects-ct

Abstract : X-ray computed tomography (CT) is a powerful tool for noninvasive imaging of time-varying objects. In thepast, methods have been proposed to reconstruct images from continuously changing objects. For discretely orstructurally changing objects, however, such methods fail to reconstruct high quality images, mainly becauseassumptions about continuity are no longer valid. In this paper, we propose a method to reconstruct structurallychanging objects. Starting from the observation that there exist regions within the scanned object that remainunchanged over time, we introduce an iterative optimization routine that can automatically determine these regionsand incorporate this knowledge in an algebraic reconstruction method. The proposed algorithm was validated onsimulation data and experimental ?CT data, illustrating its capability to reconstruct structurally changing objects moreaccurately in comparison to current techniques.

Title :An Automatic Graph-Based Approach for Artery/Vein Classification in Retinal ImagesLanguage : Java

Project Link : http://kasanpro.com/p/java/an-automatic-graph-based-artery-vein-classification-retinal-images

Abstract : The classification of retinal vessels into artery/vein (A/V) is an important phase for automating thedetection of vascular changes, and for the calculation of characteristic signs associated with several systemicdiseases such as diabetes, hypertension, and other cardiovascular conditions. This paper presents an automaticapproach for A/V classification based on the analysis of a graph extracted from the retinal vasculature. The proposedmethod classifies the entire vascular tree deciding on the type of each intersection point (graph nodes) and assigningone of two labels to each vessel segment (graph links). Final classification of a vessel segment as A/V is performedthrough the combination of the graph-based labeling results with a set of intensity features. The results of thisproposed method are compared with manual labeling for three public databases. Accuracy values of 88.3%, 87.4%,and 89.8% are obtained for the images of the INSPIREAVR, DRIVE, and VICAVR databases, respectively. Theseresults demonstrate that our method outperforms recent approaches for A/V classification.