Reconocimiento y localización de instrumental médico usando análisis automatizado de imágenes

Authors

  • Humberto Sossa
  • Roberto A. Vázquez
  • Ricardo Barrón

Abstract

 

In this work, we propose a simple but effective methodology for the recognition and localization of medical instruments from images of then. In the first stage of learning, for each instrument, we get an invariant description in the presence of image transformations such as translations and rotations. The set of descriptions is used to train several classifiers to verify their performance. During the test stage, an image containing one or more medical instruments is processed to get as a result the identity of each object in the image. For localization, for each instrument we get the coordinates of its centroid, the value of the angle of its major axis as well as the coordinates of taking holes, in the case of clamps. The proposal is proven with a set of images, specially designed for such intention.

Downloads

Download data is not yet available.

Published

2005-10-05

How to Cite

Sossa, H., Vázquez, R. A., & Barrón, R. (2005). Reconocimiento y localización de instrumental médico usando análisis automatizado de imágenes. Revista Mexicana De Ingenieria Biomedica, 26(2), 75–85. Retrieved from http://rmib.mx/index.php/rmib/article/view/253

Issue

Section

Research Articles

Dimensions Citation