Computer Vision News - September 2018

10 Computer Vision News Medical Segmentation by RSIP Vision Every month, Computer Vision News reviews a successful project. Our main purpose is to show how diverse image processing techniques contribute to solving technical challenges and physical difficulties. This month we review RSIP Vision’s Deep Learning solution for a Fully Automated Intracranial Hemorrhage and Edema Segmentation . RSIP Vision’s engineers can assist you in countless application fields. Project Blood vessel can erupt inside the brain, causing internal bleeding. This condition (called ICH or IntraCranial Hemorrhage ) can be deadly when it is not treated correctly and timely. Some brain hemorrhages come with cerebral edema : accumulating fluid in intracellular or extracellular spaces within the brain. ICH is generally diagnosed using MRI or CT scans . However, edema appears as a subtle dark area around the hemorrhage; it is therefore difficult to detect, requiring multiple scans. Successful segmentation can be achieved by expert computer vision techniques, such as performed by RSIP Vision for all fully automated medical segmentations : these techniques may be classical or built on deep learning state-of-the-art procedures such as RSIP Vision's CNN-based solutions . Watch the video above to find out more details about this procedure. It is rare to find properly annotated datasets in sufficient quantity for training. Instead of recurring to a huge manual annotation task, experts at RSIP Vision prefer to augment the dataset using an optimized semi- automatic segmentation task, by which classical computer vision techniques ( superpixels , graph cuts and more) enable a human expert to easily annotate a training dataset, subsequently used to train a small neural network. Once these tasks are completed, new techniques are called by RSIP Vision's engineers to train larger and better neural networks. The result is that RSIP can provide its clients with a fully automated medical segmentation for many types of Intracranial Hemorrhage, during which running time is constant . This Artificial Intelligence approach can be called hybrid and is most suited when training datasets are not sufficient to use the most effective deep learning methods available today. This method proposed by RSIP Vision has provided an optimal solution to many clients and may solve your own medical segmentation problem. Contact RSIP Vision today and find out how its engineers will find the best solution for your case. A fully automated medical segmentation for many types of Intracranial Hemorrhage, during which running time is constant.