Topic
Median filter
About: Median filter is a research topic. Over the lifetime, 12479 publications have been published within this topic receiving 178253 citations.
Papers published on a yearly basis
Papers
More filters
••
01 Jun 2017TL;DR: Experimental results showed that the approach is capable of recognizing fish species, which provides an effective way for solving recognition tasks in small sample size situations.
Abstract: Underwater target recognition is a challenging task due to the unrestricted environment of the ocean. With large datasets, deep learning methods have been applied with great success to the image recognition of objects in the air. However, it has been observed that deep neural networks (DNNs) easily suffer from overfitting with small samples. Underwater image acquisition always requires much manpower and costs a lot, which makes it difficult to obtain enough sample images for training DNNs. Besides, images captured by underwater cameras are usually deteriorated by noise. Taking live fish recognition as an example, we proposed a framework for underwater image recognition in small sample size situations. First, a novel improved median filter was utilized to suppress noise of fish images. Then, a convolutional neural network was employed and pre-trained with images from the world's largest image recognition database-ImageNet. Finally, preprocessed fish images were used to fine tune the pre-trained neural network and test the classification performance. Experimental results showed that the approach is capable of recognizing fish species, which provides an effective way for solving recognition tasks in small sample size situations.
51 citations
••
12 Apr 1990TL;DR: In this paper, two methods for decreasing variation due to additive noise into an image are discussed based on Singular Values Decomposition (SVD) of given image matrix: • the singular values take the meaning of the dispersion coefficients, and the Image reconstruction by part of the basis functions leads to entropy minimization of the image, guaranteeing minimization the least square error.
Abstract: Two methods for decreasing variation due to additive noise into an image are discussed. Both methods are based on Singular Values Decomposition (SVD) of given Image matrix: • The singular values take the meaning of the dispersion coefficients, and the Image reconstruction by part of the basis functions leads to entropy minimization of the image, guaranteeing minimization of the least-squares error. The sharing criterion is used by the first method to extract the most significant coefficients. • Another discussed method Is a filter fitting to the singular value spectrum of a noisy matrix. In the case of known noise distribution the filter is noise matched.
51 citations
••
TL;DR: Experimental results demonstrate the superiority of the proposed method in comparison with some of the existing methods for iris localization based on image statistics.
51 citations
••
TL;DR: The Sobel operator was found to be superior to the Roberts operator in edge enhancement and a theoretical explanation for the superior performance was developed based on the concept of analyzing the x and y Sobel masks as linear filters.
Abstract: Reference is made to the Sobel and Roberts gradient operators used to enhance image edges. Overall, the Sobel operator was found to be superior to the Roberts operator in edge enhancement. A theoretical explanation for the superior performance of the Sobel operator was developed based on the concept of analyzing the x and y Sobel masks as linear filters. By applying pill-box, Gaussian, or median filtering prior to applying a gradient operator, noise was reduced. The pill-box and Gaussian filters were more computationally efficient than the median filter with approximately equal effectiveness in noise reduction. >
51 citations
•
06 May 2005
TL;DR: In this paper, a multi-dimensional image is acquired for a first time step t; the acquired image is normalized and sampled, and then segmented into target and background pixel sets.
Abstract: Improved apparatus and methodology for image processing and object tracking that, inter alia, reduces noise. In one embodiment, the methodology is applied to moving targets such as missiles in flight, and comprises processing sequences of images that have been corrupted by one or more noise sources (e.g., sensor noise, medium noise, and/or target reflection noise). In this embodiment, a multi-dimensional image is acquired for a first time step t; the acquired image is normalized and sampled, and then segmented into target and background pixel sets. Intensity statistics of the pixel sets are determined, and a prior probability image from a previous time step smoothed. The smoothed prior image is then shifted to produce an updated prior image, and a posterior probability image calculated using the updated prior probability. Finally, the position of the target is extracted using the posterior probability image. A tracking system and controller utilizing this methodology are also disclosed.
51 citations