Author/Editor		Klopčar, N; Kerševan, K; Lenarčič, J; Valenčič, V; Pernuš, F
Title		Avtomatiziran postopek določanja tipov mišičnih vlaken
Translated title		Semiautomatic muscle fiber classification
Type		članek
Source		In: Zajc B, editor. Zbornik 9. elektrotehniške in računalniške konference ERK 2000. Zvezek B. Računalništvo in informatika, umetna inteligenca, robotika, razpoznavanje vzorcev, biomedicinska tehnika, močnostna elektrotehnika, didaktika, študentski članki; 2000 sep 21-23; Portorož. Ljubljana: Slovenska sekcija IEEE,
Publication year		2000
Volume		str. 265-8
Language		slo
Abstract		The diversity of skeletal muscles, which is reflected by the heterogeneity and spatial arrangement of their individual fibers, enables numerous movements of different velocities, forces, and endurances. The heterogeneity of muscle fibers can be assessed by histochemical techniques, which enable the classification of muscle fibres into four different types, i.e. type 1, 2a, 2b, and 2c. Manual classification is commonly carried out by following the fibers trough three serial transverse muscle slices, in which myofibrillar actomyosin adenosine triphosphatase (ATPase) activity is demonstrated at pH 9.4, pH 4.6, and pH 4.3, respectively, and by evaluating and combining the corresponding histochemical reactions. The aim of this study was to automate the classification of muscle fibers and thus increase the speed, accuracy, and reproducibility of manual classification. For this purpose, the corresponding images of muscle fibers were firstly corrected for shading and registered. Secondly, the positions of muscle fibers were determined manually in one image, what enabled automatic evaluations of corresponding histochemical reactions in all images. Finally, the reactions were mapped into a 3D space in which the parametric (k-means) and nonparametric (valley-seeking) clustering methods were implemented to distinguish between the reactions and thus classify the fibers. The results of the proposed semiautomatic classification show that both clustering methods are effective in distinguishing between type 1 and type 2 fibers but less so in distinguishing between types 2a, 2b and 2c.
Descriptors		MUSCLE, SKELETAL MUSCLE FIBERS IMAGE PROCESSING, COMPUTER-ASSISTED MIDDLE AGE AGED HISTOCYTOCHEMISTRY MYOSIN ATPASE HYDROGEN-ION CONCENTRATION MYOSIN SUBFRAGMENTS