Shape-from-focus (SFF) is extensively used in imageprocessing for obtaining shape-maps using a sequence of images of a scene captured from same view point with different camera focus settings. Many focus measure oper...
详细信息
ISBN:
(纸本)1595930361
Shape-from-focus (SFF) is extensively used in imageprocessing for obtaining shape-maps using a sequence of images of a scene captured from same view point with different camera focus settings. Many focus measure operators have been proposed in the literature for SFF applications and their relative performance depends on the camera capabilities and scene properties. In this work, we propose a novel shape estimation method which fuses individual shape estimates obtained by multiple focus measures, and reconstructs an improved shape map in a patch-wise manner, within a sparse representation (SR) framework. The SR framework involves an over-complete dictionary containing shape patches, the linear combination of which is used to reconstruct each out-put shape patch. The experimental results demonstrate that the proposed approach estimates superior shape maps in both synthetic and real scenes as compared to the estimates obtained by individual focus measures. Copyright 2014 ACM.
In this paper, we present a novel method for 3D reconstruction from epipolar plane (EP) representation of images and surface fitting for multiview or lightfield images. The proposed method detects parallelograms in EP...
详细信息
ISBN:
(纸本)1595930361
In this paper, we present a novel method for 3D reconstruction from epipolar plane (EP) representation of images and surface fitting for multiview or lightfield images. The proposed method detects parallelograms in EP images using mean shift segmentation. The slopes of the parallel lines along the left and right boundaries of the parallelograms are computed with Hough transform to estimate depth. In addition we also introduce a new concept of estimating depth from multiple EP images(diagonal EP images), making the resultant depth map denser. Multi-view depth maps are then combined and surface fitting is done to obtain the final depth map. Our method uses simple techniques to construct a quality depth map without being computationally demanding. Copyright 2014 ACM.
A limitation of using the discriminant analysis techniques for dimension reduction in image classification tasks is that the number of classes is significantly smaller than the dimension of global feature vectors used...
详细信息
ISBN:
(纸本)1595930361
A limitation of using the discriminant analysis techniques for dimension reduction in image classification tasks is that the number of classes is significantly smaller than the dimension of global feature vectors used to represent the images. In such cases, the reduced dimension is restricted by the number of classes and that reduced dimension may not adequately capture the necessary discriminatory information for classification. We propose a method to perform the discriminant analysis using the concept level labels for local feature vectors extracted from the blocks of an image. The dimension reduction is carried out on the local feature vectors. We consider the indices of the clusters of local feature vectors of all the images as the unnamed concept level labels. As the number of concept level labels is larger than the number of class level labels, the reduced dimension for local feature vectors can be higher than the number of class level labels. We consider the Gaussian mixture model based classifiers and support vector machine based classifiers for image classification using the set of local feature vectors representation of images. Results of experimental studies on image classification for MIT-8 and Vogel-6 image datasets demonstrate the effectiveness of proposed concept level discriminant analysis techniques for dimension reduction. Copyright 2014 ACM.
Long range surveillance videos are often distorted by random perturbations of the optical pathways caused by atmospheric turbulence. While humans can easily perceive and separate the moving objects from the turbulent ...
详细信息
ISBN:
(纸本)1595930361
Long range surveillance videos are often distorted by random perturbations of the optical pathways caused by atmospheric turbulence. While humans can easily perceive and separate the moving objects from the turbulent motion, it is still a challenge for vision systems. In this paper, we present a near real-time approach to detecting moving objects in the presence of turbulence. As a first step, our method learns the dynamics of the latent turbulence and extracts an approximate foreground. Statistical analysis of the foreground object properties is used to eliminate noise due to turbulence preserving only the true moving objects. Our approach also results in a stable background with minimal turbulence. We tested our method on a wide range of scenarios corrupted by various levels of turbulence. Compared with state of the art, our results are quite promising. Copyright 2014 ACM.
Motion blur is a common phenomenon in the presence of camera shake or object motion. In this paper, we deal with the challenging situation of underwater imaging. Specifically, we assume a static camera looking vertica...
详细信息
ISBN:
(纸本)1595930361
Motion blur is a common phenomenon in the presence of camera shake or object motion. In this paper, we deal with the challenging situation of underwater imaging. Specifically, we assume a static camera looking vertically down-wards at a scene but through a flowing water surface. The source of motion blur is due to the dynamic medium between the scene and the camera. Under reasonable assumptions, we establish that the motion blur induced by commonly observed fluid flows can serve as a valuable cue for inferring the underlying depth layers of the scene. We validate our approach with synthetic and real examples. Copyright 2014 ACM.
In this paper, we propose a novel method for image texture characterization. Characterization is governed by simple perceptual variations in relative orientations in terms of either no variations present or variations...
详细信息
ISBN:
(纸本)1595930361
In this paper, we propose a novel method for image texture characterization. Characterization is governed by simple perceptual variations in relative orientations in terms of either no variations present or variations present as row specific, column specific or diagonal specific. This generalization is obtained by modeling the input as a whole or image blocks depending on the broader or narrow coverage respectively. Most of the texture characterization is done either keeping a specific domain (synthetic or natural images specific to a category) or is application specific (segmentation on specific benchmark dataset or image retrieval). Contrary to this, our method is not biased towards any domain or application;rather it acts as a pre-processing step for guiding towards locating both non-textural and orientation specific textural image blocks. The proposed method quantifies the texture-tonal characterization of an image or image blocks using statistical ANOVA grading system. Once the grading for abstraction is assigned for both image as a whole and also for image blocks, the decision as to which higher-level algorithms need to be implemented on which block will become easier. The proposed method can be considered to be a three stage process - progressive sampling, image partitioning in blocks, ANOVA analysis and grading. Copyright 2014 ACM.
In this paper, we propose a novel image completion method using transform domain patch approximation method and kd-tree based nearest neighbor field (NNF) computation in multiscale fashion. In NNF, two important proce...
详细信息
ISBN:
(纸本)1595930361
In this paper, we propose a novel image completion method using transform domain patch approximation method and kd-tree based nearest neighbor field (NNF) computation in multiscale fashion. In NNF, two important processes are initialization of image target region and candidate patch searching method. Most of the previous techniques choose random initialization with arbitrary source image pixels or garbage values. It may misguide to image completion process and allow to select the bad candidate patches. We solve the problem using higher order singular value decomposition (HOSVD). It smoothly generates information in the target region enhancing the edge sharpness which helps to complete image structure quite successfully. It also preserves texture color in the target region. To overcome the problem of patch searching, we introduce robust kd-tree search method in our patch approximation step. Our experiment and analysis shows that the proposed method can be applied to the various types of image editing tools for natural images. Copyright 2014 ACM.
Video interest points, in combination with local appearance descriptors, are used for human action recognition. Most of the previously proposed video interest point detectors are straightforward extensions of some ima...
详细信息
ISBN:
(纸本)1595930361
Video interest points, in combination with local appearance descriptors, are used for human action recognition. Most of the previously proposed video interest point detectors are straightforward extensions of some image interest point detector or the other. These methods treat the temporal dimension (inter-frame) similar to the spatial dimensions (intra-frame). We argue that certain unique properties of the temporal dimension beg a different treatment. We propose an interest point detector based on vector calculus of optical flow to take advantage of the unique properties of the temporal dimension. Compared to previously proposed methods, the proposed method exhibits higher repeatability (robustness) and lower displacement (stability) of interest points under two common video transformations tested-video compression and spatial scaling. It also shows competitive action recognition performance when paired with appropriate feature descriptors in a bag of features model. Copyright is held by the authors.
We propose an example-based super-resolution (SR) framework, which uses a single input image and, unlike most of the SR methods does not need an external high resolution (HR) dataset. Our SR approach is based in spars...
详细信息
ISBN:
(纸本)1595930361
We propose an example-based super-resolution (SR) framework, which uses a single input image and, unlike most of the SR methods does not need an external high resolution (HR) dataset. Our SR approach is based in sparse representation framework, which depends on a dictionary, learned from the given test image across different scales. In addition, our sparse coding focuses on the detail information of the image patches. Furthermore, in the above process we have considered non-local combination of similar patches in the input image, which assist us to improve the quality of the SR result. We demonstrate the effectiveness of our approach for intensity images as well as range images. Contemplating the importance of edges in images of both these modalities, we have added an edge preserving constraint that will maintain the continuity of edge related information to the input low resolution image. We investigate the performance of our approach by rigorous experimental analysis and it shows to perform better than some state-of-the-art SR approaches. Copyright 2014 ACM.
This paper proposes a novel algorithm for object tracking using Approximate Nearest Neighbour Fields (ANNF). ANNF maps have been previously used to address several problems like denoising, image completion, re-targeti...
详细信息
ISBN:
(纸本)1595930361
This paper proposes a novel algorithm for object tracking using Approximate Nearest Neighbour Fields (ANNF). ANNF maps have been previously used to address several problems like denoising, image completion, re-targeting and medical image analysis. In this paper, we deal with the challenging problem of visual object tracking, using patch flow. The proposed method uses FeatureMatch to find patch corre- spondence between successive frames, enabling the tracker to find the best match for the object in the next frame. Based on the flow, each patch is labeled as either foreground or background. The proportion of FG/BG/border patches contributing to each pixel determines its final label. We show that objects can be successfully tracked across videos, under challenging conditions such as scale variations, illumination changes and occlusion using the proposed technique. Copyright 2014 ACM.
暂无评论