System and method for continuous stroke word-based text input

• US 20040140956A1
• Filed: 01/16/2003
• Published: 07/22/2004
• Est. Priority Date: 01/16/2003
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Swype, Inc. (Microsoft Corporation)
• Inventors: Kushler, Clifford A., Marsden, Randal J.
• Granted Patent: US 7,098,896 B2
• Time in Patent Office: Days
• US Class Current: 345/168
• CPC Class Codes: G06F 2203/04805 Virtual magnifying lens, i.e. window or frame movable on top of displayed information to enlarge it for better reading or selection
• CPC Class Codes: G06F 2203/04806 Zoom, i.e. interaction techniques or interactors for controlling the zooming operation
• CPC Class Codes: G06F 3/0236 using selection techniques to select from displayed items
• CPC Class Codes: G06F 3/04883 for inputting data by handwriting, e.g. gesture or text
• CPC Class Codes: G06F 3/04886 by partitioning the display area of the touch-screen or the surface of the digitising tablet into independently controllable areas, e.g. virtual keyboards or menus

Abstract

Many portable electronic devices are designed to utilize only a touch-screen for text input, generally using some form of stylus to contact the screen. Such devices generally input text using some form of handwriting recognition, which is slow and often inaccurate, or an on-screen keyboard, which essentially requires the user to perform “one-finger” typing, often on a reduced-size keyboard. The Continuous Stroke Word-Based Text Input System allows someone to use a small on-screen keyboard to quickly enter words by drawing a continuous line that passes through or near the keys of each letter in a word in sequence without lifting the stylus (similar to a children'"'"'s connect-the-dots drawing). The user traces an input pattern for a word by contacting the keyboard on or near the key of the first letter of the word, then tracing through each letter in sequence, lifting the stylus from the screen upon reaching the last letter. In one preferred embodiment, the user traces a small circle around each double letter that occurs in the word to reduce ambiguity. In another preferred embodiment, a database of words is organized according to the first and last letters so that only a small number of words need to be explicitly scored for each input pattern. In another preferred embodiment, the expected path length corresponding to each word is stored in the database and is compared to the actual input path length entered to further limit the number of words to be explicitly scored. The input pattern is analyzed to identify inflection points of various types, each of which has a greater or lesser probability of corresponding to a letter of the word being input. Words are scored according to the average distance from each letter to the nearest inflection point (or to the nearest point of the traced line if there are more letters in the word than detected inflection points in the input pattern), where each such distance is weighted according to the probability associated with the type of the inflection point from which the distance is measured. In one preferred embodiment, the average weighted distance for a word is further weighted by the frequency of use associated with the word. The method of the present invention enables the system to analyze each input stroke pattern and present one or more corresponding words for selection by the user, wherein the default word match (deemed the most probable match by the system) is automatically output by the system unless the user explicitly selects an alternate word choice.

Claims

1. A method of inputting alphabetic text to an electronic device having a virtual keyboard on a touch-sensitive screen, said virtual keyboard comprising a set of keys wherein each letter of the alphabet is associated with at least one key, by contacting the screen on or near the key associated with the first letter of a desired word, and moving the point of contact along a path that passes in succession through or near the keys associated with each letter of the desired word in sequence, and removing the point of contact from the screen at a position that is on or near the key associated with the last letter of said desired word, forming an input stroke pattern, comparing said input stroke pattern with a set of words stored in a database, determining a set of one or more words determined to most closely match the input stroke pattern, and offering one or more words of said determined set of words to the user for selection.

2. The method of claim 1, wherein the words in the database are stored in groups, wherein for each group of words, the first letters of all words in said group are associated with a single key, and the last letters of all words in said group are associated with a single key, said keys determining an ordered pair of keys uniquely associated with said group.

3. The method of claim 1, wherein the locations of the keys associated with each letter of a word in the database are compared with the input stroke pattern, and wherein the system determines a sequence of points along the path of the input stroke pattern, one point for each letter of the associated word wherein the points occur in the same sequence in the path as the associated letters occur in the associated word, such that the sum of the distances from each determined point along the path from each key associated with the associated letter is minimized, and wherein said sum of distances is used in calculating a metric to determine how closely the input stroke pattern matches said word.

4. The method of claim 3, wherein the input stroke pattern is recorded as a sequence of points of contact, each of which is identified as a location on a two-dimensional coordinate plane, and wherein a point of inflection is identified wherever the rate of change of the rate of change of either or both coordinates exceeds a determined threshold, and wherein said point of inflection is identified as the point where said rate of change of the rate of change reaches a local maximum, and wherein each said identified point of inflection is required to be used as one of said determined points from which the distance is measured to exactly one of the keys associated with the associated word.

5. The method of claim 4, wherein wherever the input stroke pattern transitions from being within a first horizontal row of keys to being within a second horizontal row of keys and then returns to again lie within the first horizontal row of keys and wherein no said point of inflection is initially determined within said second horizontal row of keys, at least one additional point of inflection is identified along the segment of the input stroke pattern lying within said second horizontal row of keys where the path of the input stroke pattern attains its maximum vertical distance from said first horizontal row of keys.

6. The method of claim 5, wherein said additional identified point of inflection is not required to be matched with a letter of a word in the database when comparing the keys associated with each letter of said word with the input stroke pattern.

7. The method of claim 1, wherein one or more of the words in the database are stored with an indication of the frequency of use associated with the word, and wherein said frequency of use is used in determining which of the set of one or more words determined to most closely match the input stroke pattern is most likely to be the word intended to be input by the user, and wherein said determined set of words is presented to the user in order of decreasing likelihood.