TheWhollySee · raw details

TheWhollySee enables the creation of large annotated data sets for training visual deep-learning systems, with data labeling and classification performed automatically at the time of image capture. This process allows neural networks to be trained on a substantially larger data set, thus increasing their image-analysis power and reducing error rates. The company's technology can also be used to train machines on the physics of motion, enabling them to infer dynamic properties of objects from static images and ultimately to predict the time evolution of static scenes from camera imagery.