FIELD OF INVENTION
The present invention relates to a method for compressing image with irregular
shaped ROI's (Region of Interest) for image transmission / storage. In particular
it relates to ROI based image compression on the touch screen touch sensitive
user interface (UI) with the aid of a stylus for transmission / storage of the
image.
BACKGROUND OF THE INVENTION
US Patent 7, 231, 086 discloses a method for detecting ROIs using prior
knowledge of the targets and the image resolution. The detected ROI can
conform to the shape of the target even if the target is of irregular shape.
Document US 6, 804, 405 provides a system and method for encoding and
transmission of still images with at least one ROI, without any information about
the target shape. There is no transmission of the boundary of the ROI. The
coefficients belonging to the ROI are shifted so that the minimum ROI coefficient

is larger than the largest background coefficient. In this process, a receiver can
then perform an opposite procedure and thereby obtain the ROI. By specifying
how much the coefficients need to be shifted in order to avoid sending shape
information several advantages are achieved. However, there is extra
computational burdens in the process.
US Patent 7, 315, 652 discloses an image processing apparatus for decoding
compressed image data which comprises a detector for detecting the existence
status of ROI within the compressed image data. There is no specific
information about the target. ROI analysis of each full tile increases
compositional cost as the transform coefficient shifts.
In the documents described above there is no specific things to tackle the
blocking artifacts and the methods are not very user friendly.
In document Select Focus Image - General Dynamics C4 Systems, the shape of
ROIs are selected from the given options which can be arbitrary.

The ROI based image codec codes are with more bits for ROI and less bits for
non-ROI (background). No steps are taken to tackle the effects caused by edges
/ demarcations while selecting an ROI specific things are provided to tackle
blocking artifacts and the method is more user friendy.
However in all the four documents described above the method is useful for
personal computer based application and there is no specific information about
implementation in embedded systems.
There was therefore, a need for a method that can be implemented using DSP,
FPGA / CPLD and ARM, microprocessor, micro controllers or digital computer.
SUMMARY OF THE INVENTION
The main object of the present invention is that Image Compression can be done
selectively by marking the ROI's,without prior knowledge of the targets and the
image resolution, using stylus,transmit / store it

Another object of the invention is to conform the detected ROI to the shape of
the target even if the target is of regular / irregular shape.
Yet another object of the present invention is to effectively detect ROIs in
common real world imagery that has any resolution.
A further object of the invention is to select a free hand, any shape ROI with the
aid of a stylus.
One more object of the present invention is to provide specific things to tackle
the blocking artifacts.
A still further object of the present invention is to provide a method which can be
implemented using a DSP, FPGA / CPLD and ARM, microprocessor, micro
controllers or digital computer, making the invention well suited to be
incorporated into an electronic product.

The image compression can be done selectively by marking the area in the
image to be compressed with the help of a stylus. Marking of the area (a close
loop of any regular or irregular shape) should be on a device having touch
sensitive user interface (UI). The image is compressed and saved using options
provided in the UI and the ROI detected compressed image is transmitted /
stored. The UI device can be an electronic visual display with touch screen
feature. For example, television, computers, mobile phones (especially for MMS)
and other multimedia display devices with touch sensitive features. Receiving
device, however need not be touch sensitive.
In a preferred embodiment of the present invention, it provides a method for
compressing image with irregular shaped ROIs (region of interest) for image
transmission / storage comprising the steps of: displaying an image on a touch
sensitive electronic visual display device; marking a part of the image as ROI in a
closed loop; obtaining frame size details and applying boundary detection and
extraction algorithm on the image followed by a threshold setting based
algorithm classifying the image as a ROI; compressing the data for compression
with ROI and non ROI (background) portions encoded with different bit rates;
and storing the ROI detected compressed data or transmitting said data to other
network / device.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
The invention can now be described in detail with the help of the figures of the
accompanying drawings in which
Figure l(a) shows in block diagram form the operating
system of the present invention.
Figure l(b) shows in block diagram form the aspect of retrieving
the ROI compressed image from memory storage and
displaying in accordance with the present invention.
DETAILED DESCRIPTION
In accordance with the present invention the image compression is done by
marking the still image displayed on an electronic visual display device provided
with a touch screen the ROI can be selected with the help of a stylus and
marked on the displayed image in the electronic visual display device. The

selection of ROI can be a free hand one and no prior knowledge of the target is
required and the selection should be a closed loop one. This is the merit of this
system. The sending of images can be on the network like HANA or other
networks using user friendly (GUI) graphic user interface which provides the
option of messaging to the receiver (the receiving device need not be a touch
screen featured device but the sending device should be a touch screen featured
device). This method is not sensitive to colour parameters and robust to noise
and other factors which affect negatively the quality of the image. Thus, this
method helps to use less memory and low bandwidth.
Image compression can be done using a stylus for marking on a still image in an
electronic visual display with touch screen (touch sensitive) feature based on
processor (DSP, FPGA, ARM, microprocessor, microcontroller or a digital
computer). This leads to less memory and low bandwidth requirements with
better picture quality.

The free hand, any shape of the ROI can be encoded form the conventional
rectangle and an ellipse of JPEG 2000, with the help of a stylus. The use of the
stylus makes the selection of ROI convenient. The novel features used in the
invention are in the free hand any shapes selection of ROI with the help of a
stylus. The method is implementable using DSP, FPGA, CPLD, ARM,
microprocessor, microcontroller or a digital computer. It is well suited to be into
an electronic product. In the present invention prior knowledge of the target is
not required and the picture quality is better.
The steps involved in the present invention are as follows:
An image is first displayed on an electronic visual display device which has a
touch screen feature. With the help of a stylus a part of the image displayed is
marked as ROI and this marking should be a closed loop one, which is important
to be stored and displayed with maximum information. The frame size details
are obtained and boundary detection and extraction algorithm is applied in the
image which is followed by a threshold setting based algorithm for classifying the
image ROI. Morphological function algorithm can be applied to trim down the

boundary line without decreasing its intensity. Proprietary segmentation
algorithm can be applied and when two different parts emerged in the image
conventional JPEG 2000 algorithm can be applied for compression with ROI
being enclosed of higher number of bits and non ROI (background) encoded of
lower number of bits. Proprietary blurring algorithm can be applied on the
boundary / demarcations of the ROI. Proprietary images enhancement algorithm
can be applied on the entire image before storing the image in memory or
transmitting it over HANA network or any other network, to device /devices. If
image retrieval is required from the memory storage ROI boundary detection can
be performed. The image can then be decoded, and displayed after image
enhancement operation.
As shown in Figure l(a) block 1 indicates the display of an image on a touch
screen based electronic visual display device. Block 2 indicates the use of a
stylus and the user can select the area of the image to be marked and then
marks it irrespective of any shape constraints. Block 3 indicates obtaining frame
size details and reducing the thickness of selected boundary of ROI.

Boundary related algorithm with respect to colour parameters with threshold
values can be applied as shown in block 4. Morphological operations are
performed as shown in block 5 to make the boundary sharper and fully
connected. In block 6 proprietary segmentation algorithm is applied. Block 7
indicates application of compression algorithm JPEG 2000. In block 8 the
boundary of ROI is blurred using proprietary algorithm. Block 9 indicates
application of proprietary image enhancement algorithm, before storing or
transmitting the data as shown in block 10.
When the data is stored in memory storage, it can be retrieved as shown in
image retrieval block 11 of Figure l(b). Block 12 indicates finding of ROI
boundary. Blurring of the boundary is indicated by block 13. Decoding the
image and reducing artifacts is indicated by block 14. Block 15 indicates
implementation of proprietary image enhancement algorithm. The
decompressed image can then be displayed as shown in block 16.

The required ROI is selected using stylus on an image displayed in touch screen
based electronic visual display. Thinning of boundary of ROI line is performed
maintaining maximum of its intensity. The frame size details are obtained and
boundary based proprietary algorithm is applied on the image with respect to the
frame size. Morphological operations are applied to make clear demarcation.
Proprietary segmentation algorithm is applied and compression is carried out.
The intensity of ROI boundary line is then reduced. After enhancing the image is
stored or transmitted. If stored in the same system then it can be retrieved, the
ROI boundary line deleted and the image can be decoded. The artifacts are
reduced and after enhancing the image it can be displayed. ^
The advantages of the present invention can be summarized as under:
The free hand selection of any shape of ROI gives the user the freedom to select
any part in the image without restricting to rectangular or ellipse shapes.
Marking ROI with stylus on touch screen based electronic visual display device
provides more user friendliness to the user. The method can be implemented
using DSP, FPGA / CPLD, and ARM, microprocessor, microcontroller or a digital
Computer.

This invention is well suited to be incorporated into an electronic product, which
would enhance its position in the market. Prior knowledge of the targets is not
required which eliminates the need for storing the prototypes (like in ANN's) or
the algorithms which describe them. By this program memory can be saved.
Also a better picture quality is obtained from the present invention.

WE CLAIM
1. A method for compressing image with irregular shaped ROIs (region of
interest) for image transmission / storage comprising the steps of:
- displaying an image on a touch sensitive electronic visual display
device;
- marking a part of the image as ROI in a closed loop;
- obtaining frame size details and applying boundary detection and
extraction algorithm on the image followed by a threshold setting
based algorithm classifying the image as a ROI;
- compressing the data for compression with ROI and non ROI
(background) portions encoded with different bit rates; and
- storing the ROI detected compressed data or transmitting said data
to other network / device.
2. The method as claimed in claim 1, wherein said marking of a part of the
image as ROI can be in any shape and free hand selection of the ROI is
done with the help of a stylus.

3. The method as claimed in claims 1 or 2, wherein the marked thickness of
the boundary line of the ROI is reduced while maintaining the intensity at
its maximum level.
4. The method as claimed in the preceding claims, wherein JPEG 2000
compression is applied so that ROI portion can be encoded with higher
number of bits and no ROI portion can be encoded with lower number of
bits.
5. The method as claimed in the preceding claims, wherein blurring
algorithm is applied on the ROI boundary with the help of a pixel pad for
diminishing the intensity of the boundary edge.
6. The method as claimed in the preceding claims, wherein said method can
be implemented with the help of DSP, FPGA / CPLD, ARM,
microprocessor, microcontroller or a digital computer.

7. The method as claimed in claim 1, wherein the stored ROI selected
compressed image is retrieved the ROI boundary is found.
8. The method as claimed in claim 7, wherein the ROI boundary line is
blurred the image is decoded and artifacts reduced.
9. The method as claimed in claim 8, wherein the decoded image is
enhanced by applying proprietory image enhancement step before
displaying in a display device.
10. A method for detecting ROIs for image compression and its transmission
/ storage, substantially as herein described and illustrated in the figures
of the accompanying drawings.

A method for compressing image with irregular shaped ROIs (region of
interest) for image transmission / storage comprising the steps of:
- displaying an image on a touch sensitive electronic visual display
device;
- marking a part of the image as ROI in a closed loop;
- obtaining frame size details and applying boundary detection and
extraction algorithm on the image followed by a threshold setting
based algorithm classifying the image as a ROI;
- compressing the data for compression with ROI and non ROI
(background) portions encoded with different bit rates; and
- storing the ROI detected compressed data or transmitting said data
to other network / device.