H. Mesa, F.J. Veredas, and L. Morente (Spain)
Pattern Analysis and Recognition, Medical Imaging, Ma chine Vision, Tissue Classification, Neural Networks, Pres sure Ulcer.
Pressure ulcer is a clinical pathology of localized damage to the skin and underlying tissue caused by pressure, shear o friction. Diagnosis, treatment and care of pressure ulcers involve high costs for the sanitary systems. Accurate pres sure ulcer evaluation is a critical task for optimizing the ef ficacy of treatments and care. Nurses usually evaluate each pressure ulcer by a visual inspection of the damaged tis sues, which is an imprecise manner of assessing the wound state. Current computer vision approaches do not offer a global solution to this particular problem. In this article, an approach based on artificial neural networks with super vised learning is used in the designing of a computational system for automatic tissue identification on pressure ul cer images. A mean shift procedure and a region-growing strategy are implemented for an effective region segmenta tion. Color and texture features are extracted from the seg mented regions. Multi-layer perceptrons are trained with inputs consisting of color and texture patterns, and with outputs consisting of categorical classes which are deter mined by clinical experts. Our outcomes show high perfor mance rates from a two-stage cascade tissue identification approach.
Important Links:
Go Back
