Abstract: A low bandwidth video teleconferencing system and method is disclosed. The video teleconferencing system employs novel data compression techniques by which continuous transmission of imagery at a rate of 9600 bits/second is possible. A sketch coder converts the grey scale image to be transmitted to a sketch or line drawing, which comprises an outline of the principal boundaries plus shading to represent depth. The bandwidth required for the data representing the sketch is then compressed by two-dimensional run length coding techniques which exploit interframe and interline redundancy as well as intraline redundancy to generate a binary transmission code. Other features are also provided.

Abstract: A dual mode facsimile data compression technique which possesses the advantage of symbol recognition and extended run-length coding methods. In operation, a symbol blocking operator isolates digital raster data representing valid alphanumeric characters and document symbols. The first symbol encoungered is placed in a memory or library and as each new symbol is detected, it is compared with each entry of the library. If the comparison is within a predetermined tolerance, a library identification code representing the symbol is generated and transmitted along with the symbol location coordinates. Otherwise, the new symbol is placed in the library and its digital pattern and location is transmitted. A scoring system determines which elements of the library are to be replaced by new prototypes once the library is filled. Non-isolated symbols are coded by a two-dimensional run-length coding method for transmission.

Abstract: A high speed character matcher and method for measuring the similarity between a pair of binary patterns representing two-dimensional images. The patterns are compared on a bit by bit basis and when a discrepancy is detected, it is assigned an error penalty which is based on the two-dimensional context in which the error occurred. The types of error context which are typical in the comparison of distorted but otherwise similar characters are given a lower error than those typically encountered in the comparison of dissimilar characters.