Ljupcho Panovski

Robust super-resolution based on pixel-level selectivity

In this paper, a new technique for robust super-resolution (SR) from compressed video is presented. The proposed method exploits the differences between low-resolution images at the pixel level, in order to determine the usability of every pixel in the low-resolution images for SR enhancement. Only the pixels, from the lowresolution images, that are determined to be usable, are included in the L2-norm minimization procedure. Three different usability criterions are proposed, maximum distance from the median - MDM, maximum distance from initial image - MDIM, and maximum distance from the SR estimate - MDSRE. The results obtained with real video sequences demonstrate superior quality of the resulting enhanced image in the presence of outliers and same quality without outliers when compared to existing L2-norm minimization techniques. At the same time, the proposed scheme produces sharper images as compared to L1-norm minimization techniques.

An adaptive POCS algorithm for compression artifacts removal

An effective compression artifacts removal algorithm is proposed based on the theory of projections onto convex sets (POCS). It includes a block classification procedure, a ringing detection procedure, prediction of the spatial distribution of the quantization errors and estimation of the visibility of the compression artifacts. Information gained from both the spatial and transform domains, is incorporated into adaptive projections. Experiments performed on JPEG-compressed images, demonstrate the effectiveness of the proposed algorithm in suppressing both blocking and ringing artifacts, as well as the ability of the algorithm to preserve the image sharpness.

Fully automated exposure fusion algorithm for mobile platforms

In this paper, an automated algorithm for fusion of differently exposed images is proposed. The algorithm is based on Gaussian/Laplacian pyramid decomposition of the input images. It includes optimization of the pyramid height for best quality results, and decision module for the necessity of the procedure for the recorded scene. The end-user involvement in the process of the creation of the output image is completely eliminated, making this algorithm a good choice for use on a mobile platform, as add-on software for low price mobile cameras. Experimental results show high efficiency of the algorithm and excellent visual quality of the resulting images.

A real time algorithm for exposure fusion of digital images

A real time algorithm for fusion of differently exposed images is proposed in this paper. The algorithm blends the details from two images of high dynamic range scene, acquired with different exposure values, into one output image which can be displayed on low dynamic range devices. The blending is performed in the spatial domain, using pixel by pixel approach, thus eliminating the need for expensive block processing or transform domain coding. The proposed scheme works both on grey and color images. The algorithm shows high efficiency, which make it applicable on low processing power platforms, such as mobile devices. The obtained results are visually comparable with previously published algorithms that are computationally much more expensive.

A real time global motion compensation for multi-exposure imaging algorithms

The paper presents a global motion compensation algorithm, designed to work in real time on low memory and low processing power hardware platforms, such as mobile phones. The algorithm is designed as one of the parts of a fully automated exposure fusion algorithm also intended for mobile platform. It implements translational shifts to one or both of the input images (the overexposed and the underexposed) in order to achieve their spatial alignment. Although the rotational and perspective movements are not included because of the high computational load for their estimation, the experimental results show that the implementation of this algorithm yields high visual quality to the fused results in most of the practical cases.

An adaptive deblocking algorithm for low bit-rate video

The paper presents an effective and fast postprocessing technique for blocking artifacts removal in low bit-rate video. The technique utilizes one-dimensional filtering based on the characteristics of the decoded video, without any knowledge about the coding parameters. In the first phase it performs detection of presence of the blocking artifacts, while in the second phase adaptive directional filtering is applied. The procedure includes estimation of spatial activity and prediction of the spatial distribution of the quantization errors. Experiments performed on MPEG-compressed low bit-rate video demonstrate the effectiveness of the proposed algorithm in suppressing blocking artifacts and the ability of the algorithm to preserve the image sharpness.

Using reduced motion vector set for real time motion registration in HDR imaging

In this paper, we propose a new global motion registration algorithm based on hierarchical motion estimation of a reduced motion vector set, specifically adapted for the properties of the HDR input images, and designed for implementation on handheld devices, demanding low computational complexity and memory load. The procedure is fully automatic and does not require any end-user involvement. The algorithm shows great effectiveness, correctly registering over 95% of the HDR input images captured by free hand.

Blind quality estimation of compressed sensing MRI

This paper presents a blind objective measure for visual quality of reconstructed MRI scans of various tissues. We analyze MRI data gathered using a customized k-space trajectory compressed sensing method which introduces visual artifacts. The goal of the proposed metric is to set a threshold for the visual quality of the sparse reconstruction in order to speed up the process of MRI acquisition and guarantee an image with a satisfactory quality for making the correct diagnosis. We use state-of-the-art machine learning for regression of local degradation features into local SSIM estimates which correlates well with human perception of visual quality. Experimental results show that a hard threshold can be applied for the needed quality during compressed sensing MRI acquisition in order to obtain optimal images for diagnosis.

A low cost algorithm for multi focus image fusion

In this paper, a low cost focus fusion algorithm is proposed. The algorithm is based on Laplacian pyramid decomposition of the input images. The included operations are designed for real time implementation on a hardware platforms with very limited resources, such as mobile devices. The whole process is fully automated. The experimental results show high speed and efficiency of the algorithm, while maintaining excellent visual quality of the resulting images.