Patitapaban Rath - Academia.edu (original) (raw)
Papers by Patitapaban Rath
International Journal of Circuit, Computing and Networking
The current methods for mesh denoising focus on models where all the vertices in the mesh are cor... more The current methods for mesh denoising focus on models where all the vertices in the mesh are corrupted with low noise amplitudes. Some of these prevailing techniques would not provide the facility of denoising even when a small percentage of the vertices are corrupted with noise of very high amplitude. To solve the problem, we propose an efficient mesh denoising algorithm for such models in which vertices are corrupted with both high as well as low noise amplitudes. This proposed algorithm is divided into two phases. In the first phase, the corrupted vertices are detected by means of a filter called Localized Coordinate Variation filter. In the second phase, the corrupted vertices are abolished. The results revealed in this paper depicts the claims as mentioned above.
International Journal of Computing and Artificial Intelligence
International Journal of Engineering in Computer Science
Arabian Journal for Science and Engineering
Our normal eye will become an infected eye if it comes under different diseases such as diabetic ... more Our normal eye will become an infected eye if it comes under different diseases such as diabetic retinopathy (DR), retinal detachment (RD) etc. Different algorithms and techniques of image processing and pattern recognition have been proposed for detecting and differentiating the normal eye from an infected eye. Automatic detection of different diseases of retina have been proposed and tested which helps the physicians to diagnose the disease with more accuracy. Fundus image is used for automatic detection of retinal diseases. The databases which are taken into consideration for detection of retinal diseases are DRIVE, STARE, REVIEW, MESSIDOR etc. Following a brief review for identifying retinal vessels where the techniques of image processing and machine learning taken into
In this paper, we put forth a progressive, decision-based, two-phase image denoising algorithm fo... more In this paper, we put forth a progressive, decision-based, two-phase image denoising algorithm for eliminating random-valued impulse noise from images. The manner in which this algorithm deals with noise is a completely pristine method when compared to the other existing image denoising algorithms. In the primary phase, the noise is dealt at a coarse level; in other words, the noisy pixels that are easily differentiable from the neighborhood are eliminated. In the secondary phase, fine-level image denoising is performed. In other words, the left-over fine scale noise in the detected corrupted pixels of the first phase, which cannot be straightforwardly differentiated from the surrounding pixels, is eliminated. In both the phases, separate mechanisms were followed to eliminate noise in the interior regions and edge regions. Hence, the algorithm is edge-detail preserving. Images with very high noise levels, in other words, with 70% noisy pixels were restored successfully. Speaking in terms of quantitative significant measures, the restored images in most cases were better than those of the other existing filters.
International Journal of Circuit, Computing and Networking
The current methods for mesh denoising focus on models where all the vertices in the mesh are cor... more The current methods for mesh denoising focus on models where all the vertices in the mesh are corrupted with low noise amplitudes. Some of these prevailing techniques would not provide the facility of denoising even when a small percentage of the vertices are corrupted with noise of very high amplitude. To solve the problem, we propose an efficient mesh denoising algorithm for such models in which vertices are corrupted with both high as well as low noise amplitudes. This proposed algorithm is divided into two phases. In the first phase, the corrupted vertices are detected by means of a filter called Localized Coordinate Variation filter. In the second phase, the corrupted vertices are abolished. The results revealed in this paper depicts the claims as mentioned above.
International Journal of Computing and Artificial Intelligence
International Journal of Engineering in Computer Science
Arabian Journal for Science and Engineering
Our normal eye will become an infected eye if it comes under different diseases such as diabetic ... more Our normal eye will become an infected eye if it comes under different diseases such as diabetic retinopathy (DR), retinal detachment (RD) etc. Different algorithms and techniques of image processing and pattern recognition have been proposed for detecting and differentiating the normal eye from an infected eye. Automatic detection of different diseases of retina have been proposed and tested which helps the physicians to diagnose the disease with more accuracy. Fundus image is used for automatic detection of retinal diseases. The databases which are taken into consideration for detection of retinal diseases are DRIVE, STARE, REVIEW, MESSIDOR etc. Following a brief review for identifying retinal vessels where the techniques of image processing and machine learning taken into
In this paper, we put forth a progressive, decision-based, two-phase image denoising algorithm fo... more In this paper, we put forth a progressive, decision-based, two-phase image denoising algorithm for eliminating random-valued impulse noise from images. The manner in which this algorithm deals with noise is a completely pristine method when compared to the other existing image denoising algorithms. In the primary phase, the noise is dealt at a coarse level; in other words, the noisy pixels that are easily differentiable from the neighborhood are eliminated. In the secondary phase, fine-level image denoising is performed. In other words, the left-over fine scale noise in the detected corrupted pixels of the first phase, which cannot be straightforwardly differentiated from the surrounding pixels, is eliminated. In both the phases, separate mechanisms were followed to eliminate noise in the interior regions and edge regions. Hence, the algorithm is edge-detail preserving. Images with very high noise levels, in other words, with 70% noisy pixels were restored successfully. Speaking in terms of quantitative significant measures, the restored images in most cases were better than those of the other existing filters.