VIDEO CAPTURE IN DATA INPUT SCENARIO Russian patent published in 2017 - IPC G06K9/46 

FIELD: information technologies.

SUBSTANCE: method of detecting the target fields in a physical document, comprising the following: information receiving by a processing device for identifying the data field in the physical document; receiving by the video stream processing device comprising multiple frames. Each frame comprises a part of the physical document; frame selection by the video stream processing device from multiple frames in a video stream; identification by the processing device of one or more text zones in the selected frame; processing by the processing unit of each of one or more identified text zones in the selected frame to identify data of each from one or more identified text zones of the selected frame, and for selection of data of one from one or more identified text zones that conform to a set of attributes associated with the data field; comparison by the data processing device of one from one or more selected text zones of identified frame with data of one or more text zones of the subsequent frame; update by the data processing device of one from one or more selected text zones of the selected frame, if the data of one or more text zones of the subsequent frame itself provides a better fit to a set of attributes; and providing by the processing device of the displayed field, containing data of one from one or more identified text zones for presentation in a user interface.

EFFECT: efficient identification of target fields in a physical document using video stream of a mobile device.

20 cl, 6 dwg