Research‎ > ‎

Medical Information Analysis

Multimodal image retrieval to support medical case-based scientific literature search

Our  long  term  and  foremost  goal  is  to  build  a  system  for  medical  case-based  information  retrieval  on  large  collections  of  scientific  biomedical  literature.  This  goal  encompasses  two  major  research  problems:  (1)  devising  an  effective  multimodal  representation  strategy  that  captures  the  rich  visual  and  textual  content  of  medical  cases  and  scientific  papers;  (2)  designing an implementing efficient algorithms and processing strategies that can cope with the ever growing collections  of  scientific  papers  and  biomedical  data.  This  includes  addressing  the  scalability  of  both  the  devised  strategies  and  algorithms and the underlying technological substrate supporting them. More...

Computational pathology

We address the problem of analyzing histopathology images, using computational tools, to automatically find patterns related with pathology signatures associated to healthy and abnormal tissues, which are a fundamental support for cancer diagnosis. Computational pathology is a relatively recent research area devoted to providing accurate and efficient computational methods to support quantitative detection, diagnosis, and prognosis in pathology. We present several computational and machine learning methods for efficient and effective automatic histopathology image analysis exploiting histopathology image databases for different digital pathology tasks including tumor and tissue detection, location and quantification in several cancer types.