FORM-2 THE PATENTS ACT, 1970
(39 of 1970)
& THE PATENTS RULES, 2003
PROVISIONAL
Specification
(See Section 10 and rule 13)


SEARCHING OF DIGITAL DATA

TATA CONSULTANCY SERVICES LTD.,
an Indian Company of Bombay House, 24, Sir Homi Mody Street, Mumbai 400 001 Maharashtra,
India
THE FOLLOWING SPECIFICATION DESCRIBES THE INVENTION


This invention relates to the field of searching of digital data.
Background
Availability and correctness of searchable content is fast growing especially with the penetration of Internet in general and mobiles in particular. Most of the content these days is user generated content (UGC) but still a large amount of content Is generated manually especially when its accuracy and relevant requirement is high. People go onto the field and feed in data into an electronic system; making the task of content generation both human intensive and erroneous. This not only adds to the cost of content creation but also takes longer time to create usable and searchable content.
Objectives
The main object of this invention is to provide a mechanism for automatic creation of content by building a content creation device which captures information from multiple sources in time sync and automatically generates content. The invention facilitates creation of content using multiple cues which include
• GPS Recoding Module - Captures the latitude and longitude of the current location along with the time stamp.
• Speech/Audio Recoding Module - a speech recording system which records audio of the speaker along with a time stamp
• lmage(Video Capturing Module - a video capturing module which captures the video with time stamp.
In accordance with this invention therefore there is provided
1. A system for automatic generation of robust and searchable content Comprising a platform (recording device) to create automatically and rapidly content related to businesses in a form that can be used to search for information using hand held or any other devices including mobile phones.
3

In accordance with a preferred embodiment of the invention, the platform can provide searchable information about businesses like name of the business, their location and any other information that relates to a company or business.
Still further, the system can extract information from multiple data sources including images, speech, position locaters
Typically, the system can be configured to work on different languages (Telugu, Tamil etc) using the support of a language tool or an electronic dictionary
The invention envisages a portable device which houses a GPS module, a sound recording module and a video recording module with the additionally facility to append time stamp on the input information stream.
Also envisaged is a mechanism to play video, audio and display the location by using GPS provided latitude and longitude information for purpose of reviewing captured content.
Particularly, the video analysis module is capable of identifying text images, extraction of image text (in different fonts and styles and different languages) into electronic text.
Particularly, the speech recognition module can convert recorded audio into electronic textual information (speech to text)
Particularly, the latitude, longitude to name of place converting mechanism uses existing latitude, longitude to name of place lookup database.
Also envisaged in the system is a mechanism to combine the outputs from video analysis, speech recognition and processed latitude, longitude information and uploading into a structured database.
4

The invention will now be described with reference to the accompanying drawings, in which
Figure 1 illustrates the block diagram of the solution architecture of the system of this invention; and
Figure 2 shows the block diagram of high level content extraction in the architecture of Figure 1.
Referring to the drawings, Figure 1 gives the high level view of the invention that allows recording of content. All the recording modules sit on a battery operated recoding device (which in a specific case could be a laptop) and the microphone, the video camera and the GSM module are connected to the data capturing device through a port (typically serial, USB, fast USB but not restricted to these modalities). The audio, video and GSM recording modules are assisted by time stamping module such that the different recordings modes have the same time stamp. All the information recorded by the devices are buffered and then stored on a storage device which could be connected externally (or connected internally) to the recording device.
Figure 2 illustrates the high-level architectural view of the content extraction methodology. All the stored information is first processed to extract the time stamp and this is appended to the information extracted from the audio, video and location processing modules. The audio processing module converts the recorded speech into electronic text; the video processing module uses video processing algorithms first identifies the regions having text and further processes to identify the language and then further processes to extract electronic text from the text image. The location is identified by using the time stamped latitude, longitude and identifies the location using latitude, longitude to location mapping database. The electronic texts


5

extracted from different modalities are synchronized using the time stamp and uploaded into a database.
This invention therefore provides a system enabling the creation of content (like business address and likes of it) automatically and robustly by fusing information from derived from multiple data sources. Further the present invention provides a platform that enables fast and automatic method of generating content to make it searchable from a mobile or any other device and a correlation between the content and its spatial location.
The Key Benefits and Features of the invention are
• An efficient and fast methodology of creating content. It beats the conventional way of manually acquiring data (time consuming) or through user generated content (error prone)
• Robust because information from different sources in evaluated separately and concatenated together; the time stamp gives authenticity to the electronic text extracted from different sources.
• The recoding device can be mounted inside a vehicle which can record (video) the businesses along the driven road and the person recording speaks the business name and or any information this is relevant and the latitude and longitude is recorded (GSM recording device). This enables a fast methodology of capturing information.
• An automated method can be employed to extract electronic text from different sources of information using video and speech processing algorithms. Making this method of extraction less prone to human errors.
While considerable emphasis has been placed herein on the specific structure of the preferred embodiment, it will be appreciated that many alterations can be made and that many modifications can be made in the preferred embodiment without departing from the principles of the invention. These and other changes in the preferred embodiment as well as other embodiments of the invention will be apparent to those
6

skilled in the art from the disclosure herein, whereby it is to be distinctly understood that the foregoing descriptive matter is to be interpreted merely as illustrative of the invention and not as a limitation.

ith
Dated this 19th day of August, 2008.


7