Mejdi Trimeche - Academia.edu (original) (raw)
Papers by Mejdi Trimeche
In this paper, we propose the use of order filters in the iterative process of super-resolution r... more In this paper, we propose the use of order filters in the iterative process of super-resolution reconstruction. At each iteration, order statistic filters are used to filter and fuse the error images. The signal dependent L-filter structure adjusts its coefficients to achieve edge preservation as well as maximum noise suppression in homogeneous regions. Depending on the amount of variance of the image pixels in different directional masks, the filter switches to use the orientation, which is most likely to follow the image edges. This procedure allows for the incorporation of a directional prior across the iterations. The introduction of a spatial filtering stage into the iterative process of super-resolution attempts to increase the robustness towards motion error and image outliers. Experimental results show the improvement obtained on sequences of noisy text images when motion is exactly known, and when a random motion error is introduced to simulate the real life situation of in...
Although the performance of imaging sensors is constantly improving, there are still several phys... more Although the performance of imaging sensors is constantly improving, there are still several physical and practical constraints that limit the final image quality. In this paper, we present a framework for producing a high-resolution color image directly from a sequence of images captured by a CMOS sensor that is overlaid with a color filter array. The algorithm attempts to utilize the additional temporal resolution in order to improve the demosaicing of the color data and filter the noisy and blurred image data. The method is based on iterative super-resolution that performs separately the filtering of the individual color image planes. We present experimental results using synthetic image sequence as well with real data from CMOS sensors.
In this paper, we present a hierarchical motion estimation a lgorithm that is based on adaptive L... more In this paper, we present a hierarchical motion estimation a lgorithm that is based on adaptive LMS filters. The algorithm is a n extension of an earlier work [1], which uses an adaptive 2-D L MS filter to match the intensity values while passing through th e image pixels according to a Hilbert scanning pattern, the algorit hm adapts the corresponding set of FIR coefficients. The peak value in t he resulted coefficient distribution points to the localized dis placement that happens between two consecutive frames. We extend the a lgorithm to use mirrored scanning and we apply it hierarchicall y cross diadic spatial resolutions. The obtained displacement at e ach resolution level is mapped to the next level, which reduces the se arch area and improves the precision of the matching process. The algorithm is particularly useful for tracking slowly varying mo ti n, such as affine or rotational motion even in the presence of noise. W e also show an example application for motion compensat...
In this paper, we present a novel approach for describing and estimating similarity of shapes. Th... more In this paper, we present a novel approach for describing and estimating similarity of shapes. The target application is content-based indexing and retrieval over large image databases. The shape feature vector is based on the efficient indexing of high curvature (HCP) points which are detected at different levels of resolution of the wavelet transform modulus maxima decomposition. The scale information,
2000 10th European Signal Processing Conference, 2000
In this paper, we present a hierarchical motion estimation algorithm that is based on adaptive LM... more In this paper, we present a hierarchical motion estimation algorithm that is based on adaptive LMS filters. The algorithm is an extension of an earlier work [1], which uses an adaptive 2-D LMS filter to match the intensity values while passing through the image pixels according to a Hilbert scanning pattern, the algorithm adapts the corresponding set of FIR coefficients. The peak value in the resulted coefficient distribution points to the localized displacement that happens between two consecutive frames. We extend the algorithm to use mirrored scanning and we apply it hierarchically across diadic spatial resolutions. The obtained displacement at each resolution level is mapped to the next level, which reduces the search area and improves the precision of the matching process. The algorithm is particularly useful for tracking slowly varying motion, such as affine or rotational motion even in the presence of noise. We also show an example application for motion compensated sharpening of video frames using the proposed LMS filtering without explicit computation of the displacement vectors.
Internet Multimedia Management Systems, 2000
ABSTRACT In this paper, we present a novel approach for describing and estimating similarity of s... more ABSTRACT In this paper, we present a novel approach for describing and estimating similarity of shapes. The target application is content-based indexing and retrieval over large image databases. The shape feature vector is based on the efficient indexing of high curvature (HCP) points which are detected at different levels of resolution of the wavelet transform modulus maxima decomposition. The scale information, together with other topological information of those high curvature points are employed in a sophisticated similarity algorithm. The experimental results and comparisons show that the technique isolates efficiently similar shapes from a large database and reflects adequately the human similarity perception. The proposed algorithm also proved efficient in matching heavily occluded contours with their originals and with other shape contours in the database containing similar portions.© (2000) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.
Although the performance of imaging sensors is constantly improving, there are still several phys... more Although the performance of imaging sensors is constantly improving, there are still several physical a nd practical constraints that limit the final image quality. In this paper, we present a framework for producing a high-resolution color image directly from a sequence of images captured by a CMOS sensor that is overlaid with a color filter array. The algorithm attempts to
A novel deconvolution technique for blurred observations corrupted by signal-dependent noise is p... more A novel deconvolution technique for blurred observations corrupted by signal-dependent noise is presented. Deblurring is performed with a transform-domain inverse-Þltering applied locally, on a sliding block of adaptively-selected pointwise varying size. Simulation results demonstrate a good quality of the proposed method, which is versatile and which can be easily combined,with other transform-domain processing. 1. INTRODUCTION AND MOTIVATION In this
The main idea of image restoration in the blur space is first to obtain a sequence of blurred ima... more The main idea of image restoration in the blur space is first to obtain a sequence of blurred images using a set of known point spread functions. Extrapolation of this sequence of images with respect to the blur parameter then gives the restored image. Usually, blur space restoration is done in a non-iterative manner and the amount of de-blurring is a parameter of the algorithm. In this paper, an iterative blur space restoration algorithm is proposed. Because of a simple stopping rule, the de-blurring parameter does not need to be predefined. Moreover, the proposed method contain a regularization procedure at pixel level that prevents edge overshooting. Results showing the improved performance of the proposed method, as opposed to the global methods, are presented.
In this paper, we propose the use of order filters in the iterative process of super-resolution r... more In this paper, we propose the use of order filters in the iterative process of super-resolution reconstruction. At each iteration, order statistic filters are used to filter and fuse the error images. The signal dependent L-filter structure adjusts its coefficients to achieve edge preservation as well as maximum noise suppression in homogeneous regions. Depending on the amount of variance of the image pixels in different directional masks, the filter switches to use the orientation, which is most likely to follow the image edges. This procedure allows for the incorporation of a directional prior across the iterations. The introduction of a spatial filtering stage into the iterative process of super-resolution attempts to increase the robustness towards motion error and image outliers. Experimental results show the improvement obtained on sequences of noisy text images when motion is exactly known, and when a random motion error is introduced to simulate the real life situation of in...
Although the performance of imaging sensors is constantly improving, there are still several phys... more Although the performance of imaging sensors is constantly improving, there are still several physical and practical constraints that limit the final image quality. In this paper, we present a framework for producing a high-resolution color image directly from a sequence of images captured by a CMOS sensor that is overlaid with a color filter array. The algorithm attempts to utilize the additional temporal resolution in order to improve the demosaicing of the color data and filter the noisy and blurred image data. The method is based on iterative super-resolution that performs separately the filtering of the individual color image planes. We present experimental results using synthetic image sequence as well with real data from CMOS sensors.
In this paper, we present a hierarchical motion estimation a lgorithm that is based on adaptive L... more In this paper, we present a hierarchical motion estimation a lgorithm that is based on adaptive LMS filters. The algorithm is a n extension of an earlier work [1], which uses an adaptive 2-D L MS filter to match the intensity values while passing through th e image pixels according to a Hilbert scanning pattern, the algorit hm adapts the corresponding set of FIR coefficients. The peak value in t he resulted coefficient distribution points to the localized dis placement that happens between two consecutive frames. We extend the a lgorithm to use mirrored scanning and we apply it hierarchicall y cross diadic spatial resolutions. The obtained displacement at e ach resolution level is mapped to the next level, which reduces the se arch area and improves the precision of the matching process. The algorithm is particularly useful for tracking slowly varying mo ti n, such as affine or rotational motion even in the presence of noise. W e also show an example application for motion compensat...
In this paper, we present a novel approach for describing and estimating similarity of shapes. Th... more In this paper, we present a novel approach for describing and estimating similarity of shapes. The target application is content-based indexing and retrieval over large image databases. The shape feature vector is based on the efficient indexing of high curvature (HCP) points which are detected at different levels of resolution of the wavelet transform modulus maxima decomposition. The scale information,
2000 10th European Signal Processing Conference, 2000
In this paper, we present a hierarchical motion estimation algorithm that is based on adaptive LM... more In this paper, we present a hierarchical motion estimation algorithm that is based on adaptive LMS filters. The algorithm is an extension of an earlier work [1], which uses an adaptive 2-D LMS filter to match the intensity values while passing through the image pixels according to a Hilbert scanning pattern, the algorithm adapts the corresponding set of FIR coefficients. The peak value in the resulted coefficient distribution points to the localized displacement that happens between two consecutive frames. We extend the algorithm to use mirrored scanning and we apply it hierarchically across diadic spatial resolutions. The obtained displacement at each resolution level is mapped to the next level, which reduces the search area and improves the precision of the matching process. The algorithm is particularly useful for tracking slowly varying motion, such as affine or rotational motion even in the presence of noise. We also show an example application for motion compensated sharpening of video frames using the proposed LMS filtering without explicit computation of the displacement vectors.
Internet Multimedia Management Systems, 2000
ABSTRACT In this paper, we present a novel approach for describing and estimating similarity of s... more ABSTRACT In this paper, we present a novel approach for describing and estimating similarity of shapes. The target application is content-based indexing and retrieval over large image databases. The shape feature vector is based on the efficient indexing of high curvature (HCP) points which are detected at different levels of resolution of the wavelet transform modulus maxima decomposition. The scale information, together with other topological information of those high curvature points are employed in a sophisticated similarity algorithm. The experimental results and comparisons show that the technique isolates efficiently similar shapes from a large database and reflects adequately the human similarity perception. The proposed algorithm also proved efficient in matching heavily occluded contours with their originals and with other shape contours in the database containing similar portions.© (2000) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.
Although the performance of imaging sensors is constantly improving, there are still several phys... more Although the performance of imaging sensors is constantly improving, there are still several physical a nd practical constraints that limit the final image quality. In this paper, we present a framework for producing a high-resolution color image directly from a sequence of images captured by a CMOS sensor that is overlaid with a color filter array. The algorithm attempts to
A novel deconvolution technique for blurred observations corrupted by signal-dependent noise is p... more A novel deconvolution technique for blurred observations corrupted by signal-dependent noise is presented. Deblurring is performed with a transform-domain inverse-Þltering applied locally, on a sliding block of adaptively-selected pointwise varying size. Simulation results demonstrate a good quality of the proposed method, which is versatile and which can be easily combined,with other transform-domain processing. 1. INTRODUCTION AND MOTIVATION In this
The main idea of image restoration in the blur space is first to obtain a sequence of blurred ima... more The main idea of image restoration in the blur space is first to obtain a sequence of blurred images using a set of known point spread functions. Extrapolation of this sequence of images with respect to the blur parameter then gives the restored image. Usually, blur space restoration is done in a non-iterative manner and the amount of de-blurring is a parameter of the algorithm. In this paper, an iterative blur space restoration algorithm is proposed. Because of a simple stopping rule, the de-blurring parameter does not need to be predefined. Moreover, the proposed method contain a regularization procedure at pixel level that prevents edge overshooting. Results showing the improved performance of the proposed method, as opposed to the global methods, are presented.