Author/Editor		Alilou, Mehdi; Kovalev, Vassili
Title		Automatic object detection and segmentation of the histocytology images using reshapable agents
Type		članek
Vol. and No.		Letnik 32, št. 2
Publication year		2013
Volume		str. 89-99
ISSN		1580-3139 - Image Analysis & Stereology online journal (Printed version).
Language		eng
Abstract		The aim of this study is to suggest a method for automatic detection and segmentation of the target objects in the microscopic histology/cytology images. The detection is carried out by rectangular shapes then segmentation process starts utilizing flexible agents which are able to move and change their shapes according to a cost function. The agents are rectangular at the beginning then they gradually fit to the corresponding objects using a stochastic reshaping algorithm. The iterative reshaping process is controlled by a cost function and it is resulted in a finer segmentation of the target objects. The cost functional of the proposed method comprised of three terms including the prior shape, regional texture and gradient information. The experiments were carried out using a publicly available microscopy image dataset which contains 510 manually-labeled target cells. The segmentation performance of the proposed method is compared with another state of the art segmentation method. The results demonstrate satisfactory detection and segmentation performance of the proposed method.
Keywords		histocytology images object detection reshapable agents segmentation