Colour Space Selection for Unsupervised Colour Image Segementation by Analysis of Connectedness Properties

Author(s):  
L. Busin ◽  
N. Vandenbroucke ◽  
L. Macaire ◽  
J.-G. Postaire
2019 ◽  
Vol 2019 (1) ◽  
pp. 243-246
Author(s):  
Muhammad Safdar ◽  
Noémie Pozzera ◽  
Jon Yngve Hardeberg

A perceptual study was conducted to enhance colour image quality in terms of naturalness and preference using perceptual scales of saturation and vividness. Saturation scale has been extensively used for this purpose while vividness has been little used. We used perceptual scales of a recently developed colour appearance model based on Jzazbz uniform colour space. A two-fold aim of the study was (i) to test performance of recently developed perceptual scales of saturation and vividness compared with previously used hypothetical models and (ii) to compare performance and chose one of saturation and vividness scales for colour image enhancement in future. Test images were first transformed to Jzazbz colour space and their saturation and vividness were then decreased or increased to obtain 6 different variants of the image. Categorical judgment method was used to judge preference and naturalness of different variants of the test images and results are reported.


2020 ◽  
pp. 53-59
Author(s):  
Vera L. Zhbanova

A problem of qualitative determination of the results of colour image capturing by digital devices based on uniform colour systems arises during analysis of colour images. The article describes methods of determining colour difference as recommended by CIE for the CIELAB system, presents mathematical formulation of the system and its modifications, describes the programme developed for calculating colour difference for each described method, and describes testing of this programme using 14 sample colours from the Munsell Book of Colour. Three groups of red, green and blue filters with 7 filters in each were selected as the study object. The filters were selected based on possibility of their further application for determining colour shift by digital devices. Filter colour difference was studied in the groups with different standard sources of light and using three methods of calculation of this difference. Analysis of the obtained results is presented and conclusions are made on possibility to apply each method in different industrial spheres. The article may be useful for colour analysis specialists, quality inspectors and specialists in digital devices and computer vision.


Author(s):  
Martin Tabakov

This chapter presents a methodology for an image enhancement process of computed tomography perfusion images by means of partition generated with appropriately defined fuzzy relation. The proposed image processing is used to improve the radiological analysis of the brain perfusion. Colour image segmentation is a process of dividing the pixels of an image in several homogenously- coloured and topologically connected groups, called regions. As the concept of homogeneity in a colour space is imprecise, a measure of dependency between the elements of such a space is introduced. The proposed measure is based on a pixel metric defined in the HSV colour space. By this measure a fuzzy similarity relation is defined, which next is used to introduce a clustering method that generates a partition, and so a segmentation. The achieved segmentation results are used to enhance the considered computed tomography perfusion images with the purpose of improving the corresponding radiological recognition.


2016 ◽  
Vol 39 (10) ◽  
pp. 1466-1485
Author(s):  
MF Kazemi ◽  
MA Pourmina ◽  
AH Mazinan

The present research attempts to address an automated optimization-based image-embedding approach through a levels-directions decomposition framework. The subject behind the present approach is to design colour image watermarking with a focus on contourlet representation, whereas the watermarking intensity is accurately calculated via an optimization algorithm with constraint. In the approach presented here, a number of performance monitors for watermarked and logo images are realized to deal with a new fitness function in the aforementioned optimization algorithm. It is worth noting that the first performance monitoring is organized based on the peak signal-to-noise ratio and the structural similarity, whereas the second one is organized based on the normal correlation and the bit error rate, respectively. There is a scrambling module to represent the logo information in disorder, where a number of attacks are simultaneously applied to the watermarked image in order to adjust the appropriate value for watermarking intensity to realize a robust and efficient solution. The ability of a decision maker system is manually taken into account for choosing the best levels and the corresponding directions regarding the contourlet representation, and the investigated results are considered in a number of well-known colour space models including RGB, YIQ and YCbCr. The key contribution of the present research is made based on the new integration of a set of subsystems employed in colour space models under the embedding and de-embedding processes in the contourlet representation, and the watermarking intensity is acquired through the optimization algorithm with constraint to present the competitive outcomes with respect to state-of-the-art benchmarks. The procedure for extracting the information concerning the logo image from the processed watermarked image under a number of attacks is implemented through the approach proposed, whereas there is no information about the original image and the watermarking intensity to be processed.


Author(s):  
Martin Welk ◽  
Andreas Kleefeld ◽  
Michael Breuß

AbstractQuantile filters, or rank-order filters, are local image filters which assign quantiles of intensities of the input image within neighbourhoods as output image values. Combining a multivariate quantile definition developed in matrix-valued morphology with a recently introduced mapping between the RGB colour space and the space of symmetric 2 × 2 matrices, we state a class of colour image quantile filters, along with a class of morphological gradient filters derived from these.We consider variants of these filters based on three matrix norms – the nuclear, Frobenius, and spectral norm – and study their differences. We investigate the properties of the quantile and gradient filters and their links to dilation and erosion operators. Using amoeba structuring elements,we devise image-adaptive versions of our quantile and gradient filters. Experiments are presented to demonstrate the favourable properties of the filters, and compare them to existing approaches in colour morphology.


Sign in / Sign up

Export Citation Format

Share Document