Abstract: A line image detection apparatus having an image input function for inputting an image on a document. The image includes a rectangular character string and a linear line drawn along the character string. The line detection apparatus has a function for detecting a character string image from the image, a function for forming a projection image on the basis of the distance from a lower end of the character string image to each first black pixel found in the character string image in the vertical direction, and a function for comparing the distance from the lower end of the character string image to each first black pixel, with a predetermined threshold value, thereby to define a plurality of "slice 0" areas included in the projection image. The distance, in each of the "slice 0" areas, from the lower end of the character string image to each first black pixel is lower than the predetermined threshold value.

Abstract: A bar-code field detecting apparatus first differentiates, in X and Y directions, image data of a document including a bar code, the position of which is not determined. It then expands each of the differential images in the direction opposite to the direction in which the image data is differentiated. Expanded images are subjected to bar-code field detection.

Abstract: A character recognition apparatus comprises a photoelectric conversion section that converts image information including the characters to be recognized into multi-valued image information according to image density, an edge extracting circuit that creates edge image information indicating the outline of image density on the basis of the multi-valued image information from the photoelectric conversion section, and a recognition circuit that computes a distribution of a direction component so as to create a feature vector and recognizes characters on the basis of the direction feature vector.

Abstract: A binarization circuit compares the density value of the target pixel with the density value of surrounding pixels in the vicinity. A judging unit judges whether or not the target pixel is isolated from the surrounding pixels. A detection unit detects the density values of the target pixels which make the target pixels isolated by a binarization process. A calculation unit calculates the accumulated isolated pixel frequencies of each density value detected by the detecting function when the judging unit and the detecting unit have finished processing all of the original image. A determination unit determines a threshold value on the basis of the accumulated isolated pixels frequencies.

Abstract: In an apparatus for detecting an address region on postal matter, an image having address information of the postal matter is converted to a digital signal, and a region associated with the digital signal is found. The associated region is converted to rectangular data indicating a rectangular region circumscribed with the associated region. The rectangular data is subjected to predetermined processing and compared with already known information, thereby easily and exactly effecting address region judgment.

Abstract: This invention is a mail address reading apparatus including a circuit for detecting image data of a mail, a circuit for detecting marks indicating the position of an address area from the image data, and a circuit for specifying the address area of the image data based on the position indicated by the marks and reading the address of the image data in the address area. In this apparatus, the address area can be specified by the detected mark position without causing any problem even if image information other than the address is present and the address can be correctly read.

Abstract: A character reading apparatus including functions for storing size information of an object to be read, an inherent image included in the object, and position information about a read area of the object, from which information is to be read, in advance, in correspondence with each other, functions for scanning the object and obtaining an actual image, functions for measuring a size of the object, functions for comparing an original image corresponding to the size of the object, stored in the storing functions, with the actual image of the object, obtained by the scanning functions, on the basis of the measurement result obtained by the measuring functions, functions for, when the comparing functions determined that the original image coincides with the actual image, determining the position information about the read area, stored in correspondence with the actual image, as a read area of the object, and functions for reading a character from an original image of the read area determined by the read area deter

Abstract: In a pattern recognition apparatus constituted by a one-chip LSI (large scale integrated), product-sum calculations are performed between an input pattern to be recognized and a plurality of reference patterns by a product-sum circuit. The plurality of reference patterns are stored in an external memory used as a dictionary. The similarity values obtained by the product-sum calculations are subjected to floating processing in a floating circuit. The similarity values, which have undergone floating processing, are sorted in the order of larger magnitudes. The sorted similarity values are read out by a host CPU (central processing unit).

Abstract: A character reading apparatus optically reads a recording medium and stores an image information as multi-value image in the multi-value image memory (3). The image information stored in the multi-value image memory (3) is scanned in the horizontal direction (character line direction). The maximum value for the pixel data of each line is detected by maximum density value extraction unit (4). When all the scanning in the horizontal direction has been completed, maximum value data (A1) is obtained. Differential data (A2) is obtained by differentially processing this maximum value data (A1), and a domain is obtained as the height of the character line. When the height of the character line have been obtained, the image information of the multi-value image memory (3) is scanned in the vertical direction. As in the horizontal direction, the maximum values for each scan in the vertical direction are extracted by maximum density value extraction unit(4).