5 results about "Document capture" patented technology

A system and methodology is provided in which image and video injection attack attempts, which may occur in connection with document or individual verification processes, can be detected and handled. The system and methodology of the present invention, in some embodiments, employs functionalities such as haptic vibration generation and accelerometer data typically present in existing user devices to implement attack detection and response.

The invention relates to a document scanner (10) comprising a housing (12) in which a document capture device (14) is arranged. The housing (12) has a scan opening (18) for the capture area (16) of the capture device (14), wherein the scan opening (18) is surrounded by a receiving area (20) of the housing (12) for a transparent support plate (22). The support plate (22) covers the scan opening (18) and serves to support a document to be scanned. According to the invention, a polymer seal (26) surrounding the scan opening (18) is arranged between the support plate (22) and the receiving area (20), making the document scanner (10) dustproof and splashproof according to protection class IP54 and thus suitable for outdoor use.

The invention relates to the technical field of deep learning and logistics document identification, in particular to a circular packaging logistics document identification and data processing method based on deep learning. The method includes: acquiring an original document image captured on site; obtaining a service context; performing physical state analysis processing on the original document image to obtain a physical state mask and a key information area physical entropy increase index; original data fragments, optical character recognition confidence and model uncertainty are obtained; combining the physical entropy increase index of the key information region, the optical character recognition confidence coefficient and the model uncertainty to solve a coupling data credibility score; based on the coupling data credibility score, the original data fragment and the service context, obtaining credible data; and generating structured data and actionable service suggestions according to credibility scores of the credible data and the coupled data. According to the method, the robustness and the environmental adaptability of circular packaging document identification in a complex field environment are remarkably improved.

Embodiments are disclosed for automated bulk document capture. The method may include receiving an input video comprising a plurality of frames. The input video depicts a plurality of document pages to be captured. A first machine learning model is used to determine a page turn event has been depicted in the input video based at least on a first frame of the input video. A second machine learning model is used to determine that a first frame of the input video is ready for capture. An image of a document page depicted in the first frame is then captured.