Searching for Knowledge in BIG Image Data Using AI
With big data, we mean any enormous and multifaceted collection of data (texts, numbers, documents, images, videos, etc.) that cannot be analyzed by ordinary computing devices and algorithms but through artificial intelligence algorithms. Big data, due to their sheer volume and inherent variety, are extremely challenging to manage and hence difficult to understand.
One of the major fields that generate big data is the biomedical and healthcare field in general and medical imaging in particular. The latter is the focus of our research at KIMIA Lab. Images do have a special place in this regard because as two-dimensional data structures, their processing is even more challenging.
More than approximately two trillion medical images are captured worldwide each year. A large number of these images have to be stored for several years. There is a huge amount of information contained in these images and their annotations (notes on diagnosis, biopsy, treatment, etc.). Presently this colossal pool of human knowledge is going untapped. Employing machine-learning algorithms on distributed platforms may help us to overcome this barrier and to create the frontier for the 21st-century medical imaging.
The Laboratory for Knowledge Inference in Medical Image Analysis, short KIMIA Lab, has been founded with the specific mandate to extract knowledge from large medical image archives by designing smart search, classification and annotation technologies.