Sharif Digital Repository

We present a modified examplar-based inpainting method in the framework of patch sparsity. In the examplar-based algorithms, the unknown blocks of target region are inpainted by the most similar blocks extracted from the source region, with the available information. Defining a priority term to decide the filling order of missing pixels ensures the connectivity of object boundaries. In the exemplar-based patch sparsity approaches, a sparse representation of missing pixels was considered to define a new priority term. Here, we modify this representation of the priority term and take measures to compute the similarities between fill-front and candidate patches. Comparative reconstructed test... 

Automatic image annotation methods generate a list of tags for each test image and present it in a matrix structure. To achieve a more accurate annotation, we propose a method with the aim of correcting the tag list. In our method, we detect an indicator for each group of tags and use it to rectify the annotation results. To find a correct indicator, we apply a deep feature vector generated by the “AlexNet” model. Using this indicator, we determine the suitable tags for an image. The purposed method is independent of feature vector, dataset, and annotation method. It can be applied to the currently available annotation methods. Our experiments showed improvement in all annotation methods... 

Identification and minimization of error sources are important issues in experimental investigations. Mainly in micro-scale problems, precise settings should be applied to high-tech test beds to reduce disturbance and induced motion. An experimental study is conducted to assess the role of induced forces and velocity fields in a particulate system used for particle identification and separation. Two main effects caused by disturbances are sampling errors and induced motion in the channel, either on fluid or dispersed phases. Different disturbance scenarios are implemented on the test bed and then the system response is reported. In order to assess induced motion as a result of applied... 

Inpainting refers to the task of filling in the missing or damaged regions of an image in an undetectable manner. We have an image to be reconstructed in a user-defined region. We use a fast decomposition method to obtain two components of the image, namely structure and texture. Reconstruction of each component is performed separately. The missing information in the structure component is reconstructed using a structure inpainting algorithm, while the texture component is repaired by a texture synthesis technique. To obtain the inpainted image, the two reconstructed components are composed together. Taking advantage of both the structure inpainting methods and texture synthesis techniques,... 

Critical crescent moon detection is a new subject in astronomical image processing. Astronomers take photos with valuable astronomical calculations using CCD cameras and telescopes. Crescent moon might not be visible in such photos, so they must be enhanced. For this purpose, we apply common image processing methods on crescent moon images such as power operator, gamma transformation, global and local histogram equalization, POSHE, CLAHE, local enhancement, gain/offset correction, homomorphic filtering, SSR, MSR, MSRCR, wavelet and curvelet transformations. Second, we use our new methods like PB, POSP, HShistV, CurvHShistV and local brightness intensity enhancement to detect the crescent.... 

The theory of compressive sampling involves making random linear projections of a signal. Provided signal is sparse in some basis, small number of such measurements preserves the information in the signal, with high probability. Following the success in signal reconstruction, compressive framework has recently proved useful in classification, particularly hypothesis testing. In this paper, conventional random projection scheme is first extended to the image domain and the key notion of concentration of measure is closely studied. Findings are then employed to develop a 2D compressive classifier (2D-CC) for sparse images. Finally, theoretical results are validated within a realistic... 

In this paper, a new blind scaling based watermarking approach is presented. The host signal is assumed to be stationary Gaussian with first-order autoregressive model. Partitioning the host signal into two separate parts, the data is embedded in one part and the other is kept unchanged for blind parameter estimation. Driving the distribution of the decision variable we have suggested a maximum likelihood decoding algorithm which is independent of the host signal distribution and can be applied for any transform domains. The proposed algorithm is applied to both artificial Gaussian autoregressive signals as well as various test images. Experimental results confirm the independence of the... 

This paper presents a new algorithm for detecting user-selected objects in a sequence of images based on a new weighted adaptive lifting scheme transform. In our algorithm, we first select a set of coefficients as object features in the wavelet transform domain and then build an adaptive transform considering the selected features. The goal of the designed adaptive transform is to "vanish" the selected features as much as possible in the transform domain. After applying both non-adaptive and adaptive transforms to a given test image, the corresponding transform domain coefficients are compared for detecting the object of interest. We have verified our claim with experimental results on 1-D...