Introduction:

Segmentation is a fundamental process that partitions a data space into meaningful salient regions. Image segmentation essentially affects the overall performance of any automated image analysis system; thus, its quality is of the utmost importance. This is a complex process with a non-exact solution and is considered an open problem, where algorithmic development has been the backbone throughout its history. Achieving efficient and well-performing strategies determines the excellent performance of an artificial intelligence system. For this reason, the fundamental objective of this work was to obtain a set of algorithms capable of achieving optimal quality in the segmentation of images, and in particular in biomedical images, which by the physics of image formation have very particular characteristics.

Methods:

In this work, an experimental research method was used by using real images, and not simulated (through computers), which guaranteed that the developed algorithms took into account the physics of their formation and the capture mode. Therefore, all the used biomedical images were obtained from databases of Cuban healthcare centers, and from standardized international sources. In addition, the developed algorithms were fine-tuned through the quantitative-qualitative research method, selecting as true criteria the images outlined and evaluations carried out by the specialists.

Results and discussion:

The results obtained with the developed strategies, according to the specialists, evidenced their effectiveness and good performance, where it should be noted that all were original ideas (autochthonous) that met the aim of achieving optimal segmentation, being able to distinguish clearly and precisely the objects or patterns that the specialists needed to accentuate. This made it possible to carry out a subsequent, simpler analysis of the pathologies associated with these bioimages. It is concluded that the proposed algorithms are original (autochthonous) and meet the aim of achieving optimal segmentation, where one can clearly and accurately distinguish the patterns that specialists need to highlight for the subsequent analysis of images.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )