November 17, 2016

Image analysis and recognition

List of publications

Image analysis and recognition

The research efforts of the MIVIA Lab in this field are mainly focused on the definition of techniques for the analysis of biomedical images with the aim of providing the enabling technologies for the realization of Computer Aided Diagnosis (CAD) systems.
Such systems may support the physician in many ways: they can be adopted as a second reader, thus augmenting the physician¹s capabilities and reducing errors; they allow to perform a pre-selection of the cases to be examined, enabling the physician to focus his/her attention only on the most relevant cases, making it easier to carry out mass screening campaigns; they may aid the physician while he/she carries out the diagnosis; finally, they can be used as a tool for training and education of specialized medical personnel.

The applicative areas investigated by the group in the recent years are the following:

  • analysis of mammographic images: mammography is a powerful tool for early diagnosis of breast cancers. Such a diagnosis is usually obtained by using precious and rare human expertise in recognizing the presence of given  patterns and types of micro-calcifications in the mammography. The contributions provided by the MIVIA lab in this applicative area are related to the automatic detection and the classification of breast microcalcifications;
  • analysis of RMN images: the reasearch activity in this area aims at developing automatic tools to aid rheumatologists in the diagnosis of bones erosion. In particular, the group have proposed methodologies for the automatic segmentation and reconstruction of the erosion of the wrist bones;
  • analysis of Indirect Immunofluorescence (IFI) images: IIF is a diagnostic methodology, based on image analysis, that reveals the presence of autoimmune diseases by searching for antibodies in the patient serum. The research activities carried out by the group are focused on the automation of the whole diagnosis process, thus from the grading of the image intensity to the detection of the cells that are in the mitotic phase up to the classification of the staining pattern.