METHOD AND SYSTEM FOR CONTENT BASED IMAGE CATEGORIZATION
FIELD
[0001] The present disclosure relates generally to the field of image processing. More
particularly, the present disclosure relates to a method and a system for content based
image categorization.
BACKGROUND
[0002] Currently, image processing applications are used to categorize images.
Existing technique such as scale invariant features (SIFT) perform categorization by using
point detector based representation of images. However, representing multiple points
involves complex processing functions thereby imposing hardware limitations for utilizing
the technique.
[0003] Further, in images having varied subjects, multiple point detectors need to be
used to identify the subjects. Using multiple point detectors leads to a higher memory
requirement, and processing cost.
[0004] In light of the foregoing discussion there is a need for a method and system for
content based image categorization to reduce the processing time and improve the
accuracy of image categorization.
SUMMARY
[0005] Embodiments of the present disclosure described herein provide amethod and
system for content based image categorization.
2
[0006] An example of a method for content based image categorization includes
identifying one or more regions of interest from multiple images. Each image from the
images is associated with a category. The method also includes extracting multiple pixels
from the one or more regions of interest in the images. Further, the method includes
determining multiple color values for the pixels in the one or more regions of interest.
Further, the method also includes grouping of the color values in a codebook
corresponding to the category. Furthermore, the method includes indexing each pixel from
among the pixels based on the color values. Furthermore, the method also includes
creating a classifier for the color values using a support vector machine.
[0007] An example of a system for content based image categorization includes an
electronic device. The electronic device includes a communication interface for receiving
multiple images that is associated with a category. The electronic device also includes a
memory for storing information. Further, the electronic device also includes a processor for
processing the information. The processor includes an identification unit for identifying one
or more regions of interest from multiple images. Each image from among the images is
associated with a category. The processor also includes an extraction unit for extracting a
multiple pixels from the one or more regions of interest. Further, the processor includes a
determination unit for determining color values for the pixels in the regions of interest.
Further, the processor also includes a grouping unit for grouping the color values in a
codebook corresponding to the category. Furthermore, the processor includes an index unit
for indexing each pixel from among the pixels based on the color values. Furthermore, the
processor also includes a classification unit for creating a classifier for the color values
using a support vector machine.
3
BRIEF DESCRIPTION OF FIGURES
[0008] The accompanying figures, similar reference numerals may refer to identical or
functionally similar elements. These reference numerals are used in the detailed
description to illustrate various embodiments and to explain various aspects and
advantages of the present disclosure.
[0009] FIG. 1 is a block diagram of a system for content based image categorization, in
accordance with which various embodiments can be implemented;
[0010] FIG. 2a-2b is a flow chart illustrating a method for content based image
categorization, in accordance with one embodiment.
[0011] FIG. 3a-3b is an exemplary illustration of categorizing multiple images, in
accordance with one embodiment.
[0012] Persons skilled in the art will appreciate that elements in the figures are
illustrated for simplicity and clarity and may have not been drawn to scale. For example, the
dimensions of some of the elements in the figures may be exaggerated relative to other
elements to help to improve understanding of various embodiments of the present
disclosure.
DETAILED DESCRIPTION
[0013] It should be observed that method steps and system components have been
represented by conventional symbols in the figures, showing only specific details that are
relevant for an understanding of the present disclosure. Further, details that may be readily
4
apparent to person ordinarily skilled in the art may not have been disclosed. In the present
disclosure, relational terms such as first and second, and the like, may be used to
distinguish one entity from another entity, without necessarily implying any actual
relationship or order between such entities.
[0014] Embodiments of the present disclosure described herein provide a method and
system for content based image categorization.
[0015] FIG. 1 is a block diagram of a system 100 for content based image
categorization, in accordance with one embodiment. The system100 includes an electronic
device 105. Examples of the electronic device 105 include, but are not limited to, a
computer, a laptop, amobile device, a hand held device, a personal digital assistant (PDA),
and a video player.
[0016] The electronic device 105 includes a bus 110 for communicating information,
and a processor 115 coupled with the bus 110 for processing information. The electronic
device 105 also includes a memory 120, for example a random access memory (RAM)
coupled to the bus 110 for storing information required by the processor 115. The memory
120 can be used for storing temporary information required by the processor 115. The
electronic device 105 further includes a read only memory (ROM) 125 coupled to the bus
110 for storing static information required by the processor 115. A storage unit 130, for
example a magnetic disk, hard disk or optical disk, can be provided and coupled to bus 110
for storing information.
[0017] The electronic device 105 can be coupled via the bus 110 to a display 135, for
example a cathode ray tube (CRT) or liquid crystal display (LCD), for displaying
5
information. An input device 140, including various keys, is coupled to the bus 110 for
communicating information to the processor 115. In some embodiments, cursor control
145, for example a mouse, a trackball, a joystick, or cursor direction keys for
communicating information to the processor 115 and for controlling cursor movement on
the display 135 can also be present.
[0018] In some embodiments, the display 135 may perform the functions of the input
device 140. For example, consider a touch screen display operable to receive haptic input.
The user can then use a stylus to select one or more portions on the visual image
displayed on the touch screen device.
[0019] In some embodiments, the steps of the present disclosure are performed by the
electronic device 105 using the processor 115. The information can be read into the
memory 120 from a machine-readable medium, for example the storage unit 130. In
alternative embodiments, hard-wired circuitry can be used in place of or in combination with
software instructions to implement various embodiments.
[0020] The term machine-readable medium can be defined as a medium providing data
to a machine to enable the machine to perform a specific function. The machine-readable
medium can be a storage media. Storage media can include non-volatile media and volatile
media. The storage unit 130 can be a non-volatile media. The memory 120 can be a
volatile media. All such media must be tangible to enable the instructions carried by the
media to be detected by a physical mechanism that reads the instructions into the machine.
6
[0021] Examples of the machine readable medium includes, but are not limited to, a
floppy disk, a flexible disk, hard disk, magnetic tape, a CD-ROM, optical disk, punchcards,
papertape, a RAM, a PROM, EPROM, and a FLASH-EPROM.
[0022] The machine readable medium can also include online links, download links,
and installation links providing the information to the processor 115.
[0023] The electronic device 105 also includes a communication interface 150 coupled
to the bus 110 for enabling data communication. Examples of the communication interface
150 include, but are not limited to, an integrated services digital network (ISDN) card, a
modem, a local area network (LAN) card, an infrared port, a Bluetooth port, a zigbee port,
and a wireless port.
[0024] Further, the electronic device 105 includes a sampler 155 for mapping each
pixel from among the pixels to the color values using vector quantization technique. The
sampler also creates an offset for the mapped pixels, the offset corresponding to the color
values in the codebook.
[0025] In some embodiments, the processor 115 includes one or more processing units
for performing one or more functions of the processor 115. The processing units are
hardware circuitry performing specified functions.
[0026] The processor includes an identification unit 160 for identifying one or more
regions of interest from the images. Each image from among the images is associated with
a category. The processor also includes an extraction unit 165 for extractingmultiple pixels
from the one or more regions of interest. Further, the processor includes a determination
7
unit 170 for determining color values for the pixels in the one or more regions of interest.
Further, the processor also includes a grouping unit 175 for grouping the color values in a
codebook corresponding to the category. Furthermore, the processor includes an index unit
180 for indexing each pixel from among the pixels based on the color values. Furthermore,
the processor also includes a classification unit 185 for creating a classifier for the color
values using a support vector machine.
[0027] In some embodiments the communication interface receives an image to be
categorized. In some embodiments, the index unit indexes each pixel of the image based
on the color values. In some embodiments the classification unit obtains the category of the
image using the classifier.
[0028] The storage unit 130 stores the codebook corresponding to the color values.
[0029] FIG. 2a-2b is a flow chart illustrating a method for content based image
categorization, in accordance with one embodiment.
[0030] The method describes training process for a classifier and performing of
categorization based on the training. Multiple images are used during the training process.
The images are associated with one or more categories. Multiple images can be associated
with each of the categories.
[0031] The method starts at step 205.
[0032] At step 210, one or more regions of interest (ROI) are identified from the
images. Each image from the images is associated with a category from among the
categories. Based on the category of each image, multiple ROIs may be identified.
8
[0033] In an embodiment, the ROIs may be identified by a user.
[0034] At step 215, a plurality of pixels is extracted from the one or more regions of
interest in the images.
[0035] At step 220, a plurality of color values for the pixels in the one or more regions of
interest are determined. The color values are based on color models, and each color value
is represented using a color correlogram vector.
[0036] Examples of the color model can include, but are not limited to, a red green blue
(RGB) model, luma-chrominance model (YCbCr), hue saturation value (HSV) color model,
and cyan, magenta, yellow and black (CMYK) model.
[0037] For example, in the RGB model, the RGB color values are determined from the
extracted pixels. The color values are represented using a 3-dimensional (D) vector
corresponding to the R, G and B colors.
[0038] At step 225, the color values are grouped in a codebook corresponding to the
category. Each grouping corresponds to a single category that can include the color values
from the multiple ROIs.
[0039] At step 230, each pixel from among the pixels are indexed based on the color
values. Here, each pixel is mapped to the color values using vector quantization technique.
An offset is created for the mapped pixel, the offset corresponding to the correlogram
vector in the codebook.
9
[0040] For example, in the RGB color model, the offset can correspond to the 3-D
vector representing the color value.
[0041] In an embodiment, the indexing reduces the number of colors in each image and
hence size of the image is reduced.
[0042] At step 235, a classifier is created for the color values using a support vector
machine (SVM). The classifier identifies the category of images using the correlogram
vectors associated with the category. A set of parameters may be defined by the classifier
using the correlogram vectors that identifies the category of the images. The SVM
constructs a hyper plane or a set of hyper planes in a high or infinite dimensional space
that can be used for classifying the images along with the correlogram vectors.
[0043] In some embodiments, an optimization process can be performed for the
classifier using an n-fold cross validation technique.
[0044] At step 240, an image is received that needs to be categorized.
[0045] At step 245, each pixel of the image is indexed based on the color values. Each
pixel of the image is mapped to the color values using the vector quantization technique.
The offset is created for the mapped pixel, the offset corresponding to the correlogram
vector in the codebook.
[0046] At step 250, the category of the image is obtained using the classifier by
identifying the category associated with the correlogram vector.
10
[0047] In an embodiment, multiple correlogram vectors are used for obtaining the
category of the image.
[0048] At step 255, the step ends.
[0049] In some embodiments, the method can be realized using one of a linear SVM
classifier and a polynomial classifier.
[0050] FIG. 3a-3b is an exemplary illustration of categorizing multiple images, in
accordance with one embodiment.
[0051] Consider multiple of images, for example an image 305A, an image 305B, an
image 305C, an image 305D, an image 305E, an image 305F, an image 305G, an image
305H, an image 305I, an image 305J, an image 305K, an image 305L, an image 305M, an
image 305N, an image 305O, and an image 305P that need to be categorized by the
classifier. Here, the image 305B, the image 305C, an image 310H, an image 310G, an
image 315I and the image 315J are rotated by 270 degrees to the viewing angle., The
classifier has been associated with categories such as mountains, monuments, water
bodies and portraits.
[0052] Each pixel of the multiple images is indexed and correlogram vector associated
with each pixel are determined. The classifier then identifies the category associated with
the correlogram vectors of each image. The images of similar category are grouped
together and displayed. For example, the image 305A, the image 305B, the image 305C
and the image 305D are grouped as the mountain category represented by the category
325. The image 305E, the image 305F, the image 305G and the image 305H are grouped
11
as the monument category represented by the category 330. The image 305I, the image
305J, the image 305K and the image 305L are grouped as the water bodies category
represented by the category 335. The image 305M, the image 305N, the image 305O and
the image 305P are grouped as the portrait category represented by the category 340.
[0053] In the preceding specification, the present disclosure and its advantages have
been described with reference to specific embodiments. However, it will be apparent to a
person of ordinary skill in the art that various modifications and changes can be made,
without departing from the scope of the present disclosure, as set forth in the claims below.
Accordingly, the specification and figures are to be regarded as illustrative examples of the
present disclosure, rather than in restrictive sense. All such possible modifications are
intended to be included within the scope of the present disclosure.
12

I/We claim:

1. A method for content based image categorization, the method comprising:
identifying one or more regions of interest from plurality of images; wherein each
image from the images is associated with a category;
extracting a plurality of pixels from the one or more regions of interest in the
images;
determining a plurality of color values for the pixels in the one or more regions of
interest;
grouping the color values in a codebook corresponding to the category;
indexing each pixel from among the pixels based on the color values; and
creating a classifier for the color values using a support vector machine.
2. The method of claim 1, wherein the indexing comprises:
mapping each pixel from among the pixels to the color values using vector
quantization technique; and
creating an offset for the mapped pixel, wherein the offset corresponds to the
color values in the codebook.
3. The method of claim 1, further comprising:
receiving an image to be categorized;
indexing each pixel of the image based on the color values; and
obtaining the category of the image using the classifier based on the indexing.
13
4. The method of claim 1, wherein the color values are based on color models.
5. The method of claim 1, wherein the color values are represented as color
correlogram vectors.
6. A system for content based image categorization, the system comprising:
an electronic device comprising:
a communication interface for receiving plurality of images associated with
a category;
a memory for storing information; and
a processor for processing the information, the processor comprising:
an identification unit for identifying one or more regions of interest
from plurality of images; wherein each image from among the images is
associated with a category;
an extraction unit for extracting a plurality of pixels from the one or
more regions of interest;
a determination unit for determining color values for the pixels in
the one or more regions of interest;
a grouping unit for grouping the color values in a codebook
corresponding to the category;
an index unit for indexing each pixel from among the pixels based
on the color values; and
14
a classification unit for creating a classifier for the color values
using a support vector machine.
7. The system of claim 6 further comprising:
a sampler to map each pixel from among the pixels to the color values using
vector quantization technique and to create an offset for the mapped pixels,
wherein the offset corresponds to the color values in the codebook.
8. The system of claim 6, wherein the communication interface is further operable to
receive an image to be categorized.
9. The system of claim 6, wherein the index unit is further operable to index each pixel
of the image based on the color values.
10.The system of claim 6, wherein the classification unit is further operable to obtain a category of the image using the classifier.