Many of todays processes benefits from digitization and digitalization of written content when it comes to data aquisition for machine learning tasks. Through the use of state of the art handwriting and gesture recognition we enable flexible multi-sketch recognition in a variety of application scenarios, ranging from radiology findings to predictive maintenance for smart factories. During wirting, real-time feedback about the handwriting recognition process is provided to the user and a set of natural gestures can be used to perform, e.g., corrections or annotations intuitively.
The picture above shows our digital radiology findings form used by radiologists to create standardized reports. During writing users receive real-time feedback about the recognition process directly on the screen. Complex gestures, such as annotations, can be performed directly on the actual medical images presented inside the form. The results are then transcribed and exported both as PDF and as structured JSON / XML to be used in later machine learning tasks.
