FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003
COMPLETE SPECIFICATION
(See Section 10 and Rule 13)
Title of invention:
A SYSTEM AND METHOD FOR DETECTION AND SEGMENTATION OF TOUCHING CHARACTERS FOR OCR
Applicants
TATA Consultancy Services Limited A company Incorporated in India under The Companies Act, 1956
Having address:
Nirmal Building, 9th Floor.
Nariman Point Mumbai 400021.
Maharashtra. India
&
Indian Statistical Institute, an Indian Institute 203 Barrackpore Trunk Road: Kolkata 700108. India
The following specification particularly describes the invention and the manner in which it is to be performed.

FIELD OF THE INVENTION
The present invention in general relates to method and system for character identification. More particularly, the invention relates to a method and system for detection of touching characters in a text output obtained from videos and images.
BACKGROUND OF THE INVENTION
An Optical Character Recognition (OCR) is the mechanical or electronic conversion of scanned images of handwritten, typewritten or printed text into machine-encoded text. Touching characters are always difficult to be recognized by an OCR engine. The binarized output of a TV video frame includes various noises like salt and pepper noise which represents itself as randomly occurring white and black pixels. Thus. there is a high possibility of false detection of characters due to the presence of this type of noise.
There are various methods available for touching character segmentation of texts obtained from videos. The main limitation of the available methods is that it cannot perform well for all text rich videos in the corpus. The existing method assumes that all characters are of almost equal width and thus the candidate cutting positions are not properly obtained thereby leading to an over segmented character. The existing methods consist of a possibility of segmentation at the left/right extension portion of the characters like "T" or "E". Moreover, the inventions are related to texts obtained from videos only.
Therefore, there is a need of a suitable touching character segmentation method which would help in improving the recognition accuracy of an OCR engine by calculating the exact touching position and segmenting the touching characters at that point only. Also, the method should be capable of being applied to text output obtained from both images and videos.

OBJECTS OF THE INVENTION
It is the primary object of the invention to provide a system and method for detection of touching characters in a media.
It is another object of the invention to provide a system and method for accurately determining the candidate cut column for those of the detected touching characters.
It is yet another object of the invention to provide a system and method for segmenting the touching characters with referential boundary thereof the candidate cut column in the component.
SUMMARY OF THE INVENTION
The present invention provides a method for detection of touching characters in a media, characterized by segmentation of adjoining character spaces. The method comprises of processor implemented steps of acquiring each component of the media in a predetermined sequence, determining an aspect ratio thereof the each component and for each detected aspect ratio higher than a threshold aspect ratio performing a component investigation. The component investigation further comprises of comprises of determining a candidate touching position of characters in a plurality of geometric orientation of the characters therein the component, computing number of pixels representing text and further computing length of a longest run of pixel representing the text for each column of the component, determining a candidate cut column based on a relation between column pixel density and corresponding length thereof the column and segmenting the touching characters with referential boundary thereof the candidate cut column in the component.
The present invention also provides a system for detection of touching characters in a media, characterized by segmentation of adjoining character spaces. The system comprises of an input means configured to acquire each component of media in a

predetermined sequence and a processor configured to determine the aspect ratio thereof each component and for performing a component investigation for each detected aspect ratio higher than a threshold aspect ratio. The processor further comprises of a determination means configured to determine a candidate touching characters in a plurality of geometric orientation of the characters therein the component, a computation means configured to compute number of pixels representing text and further computing length of a longest run of pixel the text for each column of the component, an evaluation means configured to determine a candidate cut column based on a column pixel density and corresponding length thereof of the column and a segmentation means configured to segment the touching characters with referential boundary thereof the candidate cut column in the component.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 illustrates the system architecture in accordance with an embodiment of the invention.
Figure 2 illustrates an exemplary flowchart in accordance with an alternate embodiment of the invention.
DETAILED DESCRIPTION
Some embodiments of this invention, illustrating its features, will now be discussed:
The words "comprising", "having", "containing", and "including", and other forms thereof, are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items, or meant to be limited to only the listed item or items.
It must also be noted that as used herein and in the appended claims, the singular forms "a", "an", and "the" include plural references unless the context clearly dictates

otherwise. Although any systems, methods, apparatuses, and devices similar or equivalent to those described herein can be used in the practice or testing of embodiments of the present invention, the preferred, systems and parts are now described. In the following description for the purpose of explanation and understanding reference has been made to numerous embodiments for which the intent is not to limit the scope of the invention.
One or more components of the invention are described as module for the understanding of the specification. For example,, a module may include self-contained component in a hardware circuit comprising of logical gate, semiconductor device. integrated circuits or any other discrete component. The module may also be a part of any software programme executed by any hardware entity for example processor. The implementation of module as a software programme may include a set of logical instructions to be executed by the processor or any other hardware entity. Further a module may be incorporated with the set of instructions or a programme by means of an interface.
The disclosed embodiments are merely exemplary of the invention, which may be embodied in various forms.
The present invention relates to a system and a method for detection of touching characters in a media. The detection of touching characters is further characterized by segmentation of adjoining character spaces. In the very first step, an aspect ratio is calculated for each connected component. A candidate touching position of each character is determined by calculating a threshold aspect ratio for each character. Further a candidate cut column is determined based on a relation between column pixel densities and corresponding length thereof the column. Lastly, the touching characters segmented with referential boundary thereof the candidate cut column.

In accordance with an embodiment, referring to figure 1, the system (100) comprises of an input means (102) which is configured to acquire each component of media in a predetermined sequence. The system (100) further comprises of a processor (104) which processes the component of media for component investigation. The processor (104) further comprises of a determination means (106), a computation means (108), an evaluation means (110) and segmentation means (112).
In accordance with an.embodiment, still referring to figure 1. an input means (102) acquires each component of media in a predetermined sequence wherein the media includes images captured from camera, scanned images or videos. The processor (104) is further configured to determine the aspect ratio thereof each component (as shown in step 202 of figure 2).
The processor (104) further comprises of a binarization module (not shown in the figure) which binarizes the media before determining the aspect ratio. The aspect ratio is calculated as,
Aspect Ratio (A) = w/h
Where, w = width of the component and h = height of the component
The processor (104) further comprises of a determination means (106) which is configured to determine a candidate touching characters in a plurality of geometric orientation of the characters therein the component. The candidate touching characters in a component is determined by comparing the aspect ratio of that component with a threshold aspect ratio (as shown in step 204 of figure 2).
In accordance with an embodiment, by way of a specific example, the value of the threshold aspect ratio is 1. Therefore,
A component is a candidate touching character if,

Aspect Ratio (A) > 1
If the aspect ratio is not greater than 1 then the component is not considered as a candidate touching character and the next component from the input is considered for processing (as shown in step 206 of figure 2).
The processor (104) further comprises of a computation means (108) which is configured to compute number of pixels representing text (as shown in step 208 of figure 2) and the length of a longest run of the pixel representing the text (as shown in step 210 of figure 2), for each column of the component. The number of pixels representing text (cnt_txt) and the length of the longest run of the pixels representing text (longest_txt) are calculated for each column of a candidate touching character.
The processor further comprises of an evaluation means (110) which is configured to determine a candidate cut column. A feature r is defined which indicates the possibility of the ith column being a candidate cut column as,
r i = cnt_txt * longest__txt (as shown in step 212 of figure 2)
A column is marked as a candidate cut column if the following condition is satisfied:
T j+i > -{r <\T _] and T j < threshold (as shown in step 214 of figure 2)
The threshold value is obtained heuristically. By way of a specific example, the threshold value in case of Indian TV videos is 10.
The processor (104) further comprises of a segmentation means (112) which is configured to segment the touching characters with referential boundary thereof the candidate cut column in the component.

BEST MODE/EXAMPLE FOR WORKING OF THE INVENTION
The system and method illustrated for detection of touching characters in a media,
characterized by segmentation of adjoining character spaces, to be interpreted further
for a specific purpose may be illustrated by working examples stated in the following
paragraph; the process is not restricted to the said examples only:
Let us consider an image captured from a running video of a Bollywood movies
include a text showing name of a movie. When the image is captured it is affected by
a plurality of noises (shadow of other objects etc) and the characters are seemed to be
overlapped. This image if video being processed OCR (Optica) character recognition)
is first being enhanced by the above described system and method.
The aspect ratio is first determined and is compared to a threshold value (for example
1). If this aspect ratio is greater than 1, then it is processed for further investigation. In
the further investigation candidate touching position are determined in a plurality of
geometric orientations.
Further out the complete image area, the numbers of pixels which are representing
text are computed and then the length of a longest run of pixel representing the text
for each column of the component is computed.
Considering the product of the number of pixels representing text and the length of
longest run of text pixel, candidate cut column are determined. Lastly, the touching
characters are being segmented with referential boundary thereof the column before
being processed by the OCR for identification of the movie name.

WE CLAIM:
1. A method for detection of touching characters in a media, characterized by
segmentation of adjoining character spaces, the method comprising:
acquiring each component of the media in a predetermined sequence;
determining an aspect ratio thereof the each component, for each detected
aspect ratio higher than a threshold aspect ratio perform a component
investigation, the component investigation comprising:
determining a candidate touching position of characters in a plurality
of geometric orientation of the characters therein the component;
computing number of pixels representing text and further computing
length of a longest run of pixel representing the text for each column
of the component;
determining a candidate cut column based on a relation between
column pixel density and corresponding length thereof the column;
and
segmenting the touching characters with referential boundary thereof
the candidate cut column in the component.
2. The method as claimed in claim 1, wherein the media includes images captured from camera, scanned images or videos.
3. The method as claimed in claim 1, wherein the media may be first binarized before determining the aspect ratio.
4. The method as claimed in claim 1, wherein a value of threshold aspect ratio is 1.

5. The method as claimed in claim 1. wherein the relation includes a product of number of pixels representing the text and the length of longest run of pixels.
6. A system for detection of touching characters in a media, characterized by segmentation of adjoining character spaces, the system comprising:
an input means configured to acquire each component of media in a predetermined sequence;
a processor configured to determine the aspect ratio thereof each component and for performing a component investigation for each detected aspect ratio higher than a threshold aspect ratio, the processor further comprising:
a determination means configured to determine a candidate touching
characters in a plurality of geometric orientation of the characters
therein the component;
a computation means configured to compute number of pixels
representing text and further computing length of a longest run of
pixel the text for each column of the component;
an evaluation means configured to determine a candidate cut column
based on a column pixel density and corresponding length thereof of
the column; and
a segmentation means configured to segment the touching characters
with referential boundary thereof the candidate cut column in the
component.
7. The system as claimed in claim 6. wherein the media includes images captured from camera, scanned images or videos.
8. The system as claimed in claim 6, wherein a value of threshold aspect ratio is 1.

9. The system as claimed in claim 6, wherein the system further comprises of binarization module configured to binarize the media before determining the aspect ratio.