Method and system for spatial, appearance and acoustic coding of words and sentences
AssigneeDybuster AG
InventorMarkus Gross
Life of this patent
Estimating the term…
Abstract
A method encodes a word or words. A sequential input string of symbols representing a word or a plurality of words is parsed into segments. A graph is constructed having spatial levels, each level including nodes. The segments of the input string are mapped to the nodes according to the levels and attributes are assigned to the nodes according to the segments, where an entropy of the word or plurality of words is a constant times an entropy of the graph and of the nodal attributes.
Claims16
A method for training dyslexic students on improving reading skills using a computer-based system that is capable of displaying at least text from a training data set, the method comprising:
- receiving the training data set as a sequence of symbols representing one or more words from the text of the training data set;
- parsing the sequence of symbols into a set of segments, the segments representing one or more of words, syllables or letters;
- generating a tree graph data structure, in a computer-readable medium, representing a tree graph having a plurality of spatial levels, each level including nodes, wherein child nodes in the tree graph correspond to segments of the set of segments, with nodes of a particular level corresponding to words, syllables or letters;
- mapping attributes from a set of attributes to nodes of the tree graph, such that the sum of an entropy of the graph and an entropy of the set of attributes is a constant times an entropy of the one or more words;
- and
- displaying, to a dyslexic student, using a display device of the computer-based system, a representation of the sequence of symbols with the attributes visually represented, thereby providing additional display information beyond just the sequence of symbols usable to training on improving reading skills.
Dependencies are inferred from the claim text.