Method for recognizing pattern of text documents for recognizing pedestrian, in surrounding area of vehicle, involves determining code word, where sparseness coding of data matrix is formed based on all determined code words

摘要

The method involves forming a data matrix (X) from an input (x) i.e. vehicle surrounding area image. The data matrix is factorized into a filter matrix (W) and a code word matrix (H). Columns of the filter matrix are formed of basic vectors. The filter matrix and/or the code word matrix are limited with sparseness measures (sigma). The input is scanned via the filter matrix for determination of data representation (D). A code word is determined, where a sparseness coding of the data matrix is formed based on all determined code words.