04-09-2012, 01:24 PM
Image Compression: An approach using Wavelet Transform and Modified FCM
[attachment=33312]
ABSTRACT
In recent past, vector quantization has been observed as an
efficient technique for image compression. In general, image
compression reduces the number bits required to represent an
image. The main significance of image compression is that the
quality of the image is preserved. This in turn increases the
storage space and thereby the volume of the data that can be
stored. Image compression is the application of data
compression technique on digital images. Wavelet Transform
based image compression remain the most common among
diverse techniques proposed earlier. Wavelet-based image
compression provides considerable improvements in picture
quality at higher compression ratios. A moment ago Artificial
Neural Network has attained popularity in the field of image
compression. This paper proposes a technique for image
compression using modified Fuzzy C-Means (FCM) algorithm
based vector quantization (VQ). The VQ codebook is generated
by a modified FCM algorithm. The principal shortcoming of
standard FCM algorithm is that the objective function does not
think about the spatial dependence therefore it deal with image
as the same as separate points.
INTRODUCTION
Image compression is an indispensable characteristic of image
processing without which it is difficult to transmit an image of
large size through internet etc. Image compression is the
relevance of data compression on digital images. In effect, the
main purpose of image compression is to trim down redundancy
of the image data in order to be able to store or transmit data in
an efficient form. Image compression is reducing the size in
bytes of a graphics file without debasing the quality of the image
to an unacceptable level. The image compression approaches
can be divided into two methods: lossless and lossy. The
decrease in file size allows more images to be stored in a given
amount of disk or memory space.
RELATED WORK
Image compression is an essential feature of image processing
without which it is difficult to transmit an image of large size
through internet etc. This section of the paper discusses some of
the earlier work proposed on image compression using neural
networks and wavelet transform.
Khashman et al. in [8] proposed a technique for compressing the
digital image using neural networks and Haar Wavelet
transform. The aim of the work presented within the paper was
to develop an optimum image compression system using haar
wavelet transform and a neural network. With Wavelet
transform based compression, the quality of compressed images
is typically high, and the option of a perfect compression ratio is
complicated to formulate as it varies depending on the content of
the image. They proposed that neural networks can be trained to
ascertain the non-linear relationship between the image intensity
and its compression ratios in search for an optimum ratio.
Moreover their paper suggested that a neural network could be
trained to be familiar with an optimum ratio for Haar wavelet
compression of an image upon presenting the image to the
network. The method utilized Haar compression with nine
compression ratios and a supervised neural network that learns
to correlate the grey image intensity (pixel values) with a single
optimum compression ratio. Two neural networks receiving
different input image sizes are developed in their work and a
comparison between their performances in finding optimum
Haar-based compression was presented.
PROPOSED APPROACH
Wavelets for image compression
Wavelet transform make use of both the spatial and frequency
correlation of data by dilations (or contractions) and translations
of mother wavelet on the input data. It supports the
multiresolution analysis of data i.e. it can be applied to various
scales according to the details required, which allows
progressive transmission and zooming of the image without the
need of extra storage. One more encouraging feature of wavelet
transform is its symmetric behavior that is both the forward and
the inverse transform has the same complexity, building fast
compression and decompression routines. Its characteristics well
suited for image compression comprise the ability to take into
account of Human Visual System’s (HVS) characteristics, very
good energy compaction capabilities, robustness under
transmission, high compression ratio etc.
The implementation of wavelet compression method is very
similar to that of subband coding scheme: the signal is
decomposed using filter banks. The output of the filter banks is
down-sampled, quantized, and encoded. The decoder decodes
the coded representation, up-samples and recomposes the signal.
CONCLUSION
FCM is an extensively accepted clustering method and has been
widely applied for image compression. This proposed paper
introduced a modified Fuzzy C-Means algorithm for image
compression. This variation overcomes the restraint of the
standard FCM algorithm. The Vector Quantization (VQ)
codebook is generated by a modified FCM algorithm. This
proposed paper modifies the standard FCM algorithm that
integrates both the local spatial context and the non
information into the standard FCM cluster algorithm using a
novel dissimilarity index in place of the usual distance metric.
The membership value decides the compression results and
hence the membership value is evaluated by the distance
measurement. In order to estimate the performance of modified
FCM based vector quantization for image compression some
standard image set are considered.
[attachment=33312]
ABSTRACT
In recent past, vector quantization has been observed as an
efficient technique for image compression. In general, image
compression reduces the number bits required to represent an
image. The main significance of image compression is that the
quality of the image is preserved. This in turn increases the
storage space and thereby the volume of the data that can be
stored. Image compression is the application of data
compression technique on digital images. Wavelet Transform
based image compression remain the most common among
diverse techniques proposed earlier. Wavelet-based image
compression provides considerable improvements in picture
quality at higher compression ratios. A moment ago Artificial
Neural Network has attained popularity in the field of image
compression. This paper proposes a technique for image
compression using modified Fuzzy C-Means (FCM) algorithm
based vector quantization (VQ). The VQ codebook is generated
by a modified FCM algorithm. The principal shortcoming of
standard FCM algorithm is that the objective function does not
think about the spatial dependence therefore it deal with image
as the same as separate points.
INTRODUCTION
Image compression is an indispensable characteristic of image
processing without which it is difficult to transmit an image of
large size through internet etc. Image compression is the
relevance of data compression on digital images. In effect, the
main purpose of image compression is to trim down redundancy
of the image data in order to be able to store or transmit data in
an efficient form. Image compression is reducing the size in
bytes of a graphics file without debasing the quality of the image
to an unacceptable level. The image compression approaches
can be divided into two methods: lossless and lossy. The
decrease in file size allows more images to be stored in a given
amount of disk or memory space.
RELATED WORK
Image compression is an essential feature of image processing
without which it is difficult to transmit an image of large size
through internet etc. This section of the paper discusses some of
the earlier work proposed on image compression using neural
networks and wavelet transform.
Khashman et al. in [8] proposed a technique for compressing the
digital image using neural networks and Haar Wavelet
transform. The aim of the work presented within the paper was
to develop an optimum image compression system using haar
wavelet transform and a neural network. With Wavelet
transform based compression, the quality of compressed images
is typically high, and the option of a perfect compression ratio is
complicated to formulate as it varies depending on the content of
the image. They proposed that neural networks can be trained to
ascertain the non-linear relationship between the image intensity
and its compression ratios in search for an optimum ratio.
Moreover their paper suggested that a neural network could be
trained to be familiar with an optimum ratio for Haar wavelet
compression of an image upon presenting the image to the
network. The method utilized Haar compression with nine
compression ratios and a supervised neural network that learns
to correlate the grey image intensity (pixel values) with a single
optimum compression ratio. Two neural networks receiving
different input image sizes are developed in their work and a
comparison between their performances in finding optimum
Haar-based compression was presented.
PROPOSED APPROACH
Wavelets for image compression
Wavelet transform make use of both the spatial and frequency
correlation of data by dilations (or contractions) and translations
of mother wavelet on the input data. It supports the
multiresolution analysis of data i.e. it can be applied to various
scales according to the details required, which allows
progressive transmission and zooming of the image without the
need of extra storage. One more encouraging feature of wavelet
transform is its symmetric behavior that is both the forward and
the inverse transform has the same complexity, building fast
compression and decompression routines. Its characteristics well
suited for image compression comprise the ability to take into
account of Human Visual System’s (HVS) characteristics, very
good energy compaction capabilities, robustness under
transmission, high compression ratio etc.
The implementation of wavelet compression method is very
similar to that of subband coding scheme: the signal is
decomposed using filter banks. The output of the filter banks is
down-sampled, quantized, and encoded. The decoder decodes
the coded representation, up-samples and recomposes the signal.
CONCLUSION
FCM is an extensively accepted clustering method and has been
widely applied for image compression. This proposed paper
introduced a modified Fuzzy C-Means algorithm for image
compression. This variation overcomes the restraint of the
standard FCM algorithm. The Vector Quantization (VQ)
codebook is generated by a modified FCM algorithm. This
proposed paper modifies the standard FCM algorithm that
integrates both the local spatial context and the non
information into the standard FCM cluster algorithm using a
novel dissimilarity index in place of the usual distance metric.
The membership value decides the compression results and
hence the membership value is evaluated by the distance
measurement. In order to estimate the performance of modified
FCM based vector quantization for image compression some
standard image set are considered.