Image understanding solutions using the Cognitive Workbench
ExB’s proprietary computer vision specialized solutions for healthcare and life sciences use convolutional neural networks, a special form of deep learning, to identify objects or anomalies in various types of medical images, e.g., in Histology slices or MRI scans. These Networks mimic the human brain and learn from input data instead of using handcrafted features. This makes the approach very flexible and accurate.
The increasing number medical images combined with exploding richness of details of daily gen demands innovative and smart ways of image analysis.
Automating the segmentation and characterization of regions of interest minimizes errors, fastens the process and improves comparability.
Histology studies the microscopic structures of the organic tissues and has many standard applications in medicine. The analysis of the slice images is an image segmentation problem where cells or their kernels need to be recognized.
Today, these complex tasks are still predominantly performed manually by trained experts – a lengthy and expensive process which would clearly benefit from automated image analysis.
One common application of Magnetic resonance imaging (MRI) is image analysis as basis for tumor classification and growth progression tracking. The manual segmentation of MRIs is very time consuming and the standard process becomes increasingly impractical, especially for today’s large sets of data.. ExB Health delivers a deep learning based solution for MRI segmentation and classification that provides high accuracy, easy handling and fast processing.