Premium
Automatic seamless mosaicing of microscopic images: enhancing appearance with colour degradation compensation and wavelet‐based blending
Author(s) -
HSU W.Y.,
POON W.F. PAUL,
SUN Y.N.
Publication year - 2008
Publication title -
journal of microscopy
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.569
H-Index - 111
eISSN - 1365-2818
pISSN - 0022-2720
DOI - 10.1111/j.1365-2818.2008.02052.x
Subject(s) - artificial intelligence , computer vision , computer science , magnification , projection (relational algebra) , distortion (music) , wavelet , feature (linguistics) , compensation (psychology) , pattern recognition (psychology) , algorithm , psychology , amplifier , computer network , linguistics , philosophy , bandwidth (computing) , psychoanalysis
Summary In order to observe the fine details of biomedical specimens, various kinds of high‐magnification microscopes are used. However, they suffer from a limited field of view when visualizing highly magnified specimens. Image mosaicing techniques are necessary to integrate two or more partially overlapping images into one and make the whole specimen visible. In this study, we propose a new system that automatically creates panoramic images by mosaicing all the microscopic images acquired from a specimen. Not only does it effectively compensate for the congenital narrowness in microscopic views, but it also results in the mosaiced image containing as little distortion with respect to the originals as possible. The system consists of four main steps: (1) feature point extraction using multiscale wavelet analysis, (2) image matching based on feature points or by projection profile alignment, (3) colour difference adjustment and optical degradation compensation with a Gaussian‐like model and (4) wavelet‐based image blending. In addition to providing a precise alignment, the proposed system also takes into account the colour deviations and degradation in image mosaicing. The visible seam lines are eliminated after image blending. The experimental results show that the system performs well on differently stained image sequences and is effective on acquired images with large colour variations and degradation. It is expected to be a practical tool for microscopic image mosaicing.