18-06-2013, 04:52 PM
An Improved Support Vector Machine Kernel for Medical Image Retrieval System
An Improved Support.docx (Size: 473.3 KB / Downloads: 25)
ABSTRACT
Digital medical images take up most of the storage space in the medical database. Digital images are in the form of X-Rays, MRI, CT. These medical images are extensively used in diagnosis and planning treatment schedule. Retrieving required medical images from the database in an efficient manner for diagnosis, research and educational purposes is essential. Image retrieval systems are used to retrieve similar images from database by inputting a query image. Image retrieval systems extract features in the image to a feature vector and use similarity measures for retrieval of images from the database. So the efficiency of the image retrieval system depends upon the feature selection and its classification. In this paper, it is proposed to implement a novel feature selection mechanism using Discrete Sine Transforms (DST) with Information Gain for feature reduction. Classification results obtained from existing Support Vector Machine (SVM) is compared with the proposed Support Vector Machine model. Results obtained show that the proposed SVM classifier outperforms conventional SVM classifier and multi layer perceptron neural network.
INTRODUCTION
Digital images provide visual information required for diagnosis and progress in medical treatment. Image retrieval of digital medical images from archives is a challenge that is widely researched. Textual annotations of images were the basis on which images were retrieved during the early 80s [1, 2]. Semantic queries were used for retrieving images. Textual annotations of the medical images were manually done using keywords. Retrieval was based on these keywords. Manually annotating using keywords involves a huge amount of manuallabor and with the increasing volume of digital images stored, it is not feasible. Thus, the need to retrieve images based on the content of the image rather than using metadata such as keywords, and tags for efficient use of medical database. A medical retrieval system which can automatically classify images based on the features of the image and retrieve images based on query image is required. Earlier works in literature include use of visual features with text annotation for image retrieval [3, 4].
Modern radiology techniques like CT, PET, MRI, X-Rays, provide essential information required for diagnose and plan treatments to the medical professionals [5]. Thus, retrieval of similar images from the database using query image for diagnosis is required. Visual features in the query image is extracted and compared to retrieve query similar images from the archive. Features such as color, texture, shape and spatial relationship are used for classifying images. These features are compared for similarity using an image distance measure. Color is an effectively used feature in fields of dermatology [8] color is extensively used as a feature. MRI images, X-Rays are in grey scale, thus instead of color, shape and spatial relationship are used for similarity measures. Similarity measures computed from low level image features are mainly used for image retrieval. To automatically categorize medical images, data mining techniques such as decision tree, Bayesian network, Neural networks, Support vector machines are widely used [9].
DISCRETE WAVELET TRANSFORM
Introduction
The transform of a signal is just another form of representing the signal. It does not change the information content present in the signal. The Wavelet Transform provides a time-frequency representation of the signal. It was developed to overcome the short coming of the Short Time Fourier Transform (STFT), which can also be used to analyze non-stationary signals. While STFT gives a constant resolution at all frequencies, the Wavelet Transform uses multi-resolution technique by which different frequencies are analyzed with different resolutions.
A wave is an oscillating function of time or space and is periodic. In contrast, wavelets are localized waves. They have their energy concentrated in time or space and are suited to analysis of transient signals. While Fourier Transform and STFT use waves to analyze signals, the Wavelet Transform uses wavelets of finite energy.
Multi-Resolution Analysis using Filter Banks
Filters are one of the most widely used signal processing functions. Wavelets can be realized by iteration of filters with rescaling. The resolution of the signal, which is a measure of the amount of detail information in the signal, is determined by the filtering operations, and the scale is determined by upsampling and downsampling (subsampling) operations[5].
The DWT is computed by successive lowpass and highpass filtering of the discrete time-domain signal as shown in figure 2.2. This is called the Mallat algorithm orMallat-tree decomposition. Its significance is in the manner it connects the continuoustime mutiresolution to discrete-time filters. In the figure, the signal is denoted by the sequence x[n], where n is an integer. The low pass filter is denoted by G0 while the high pass filter is denoted by H0. At each level, the high pass filter produces detail information, d[n], while the low pass filter associated with scaling function produces coarse approximations, a[n].
Conditions for Perfect Reconstruction
In most Wavelet Transform applications, it is required that the original signal be synthesized from the wavelet coefficients. To achieve perfect reconstruction the analysis and synthesis filters have to satisfy certain conditions. Let G0(z) and G1(z) be the low pass analysis and synthesis filters, respectively and H0(z) and H1(z) the high pass analysis and synthesis filters respectively. Then the filters have to satisfy the following two conditions as given in [4] :
The first condition implies that the reconstruction is aliasing-free and the second condition implies that the amplitude distortion has amplitude of one. It can be observed that the perfect reconstruction condition does not change if we switch the analysis and synthesis filters.
There are a number of filters which satisfy these conditions. But not all of them give accurate Wavelet Transforms, especially when the filter coefficients are quantized. The accuracy of the Wavelet Transform can be determined after reconstruction by calculating the Signal to Noise Ratio (SNR) of the signal. Some applications like pattern recognition do not need reconstruction, and in such applications, the above conditions need not apply.
Applications
There is a wide range of applications for Wavelet Transforms. They are applied indifferent fields ranging from signal processing to biometrics, and the list is still growing. One of the prominent applications is in the FBI fingerprint compression standard. Wavelet Transforms are used to compress the fingerprint pictures for storage in their databank. The previously chosen Discrete Cosine Transform (DCT) did not perform well at high compression ratios. It produced severe blocking effects which made it impossible to follow the ridge lines in the fingerprints after reconstruction. This did not happen with Wavelet Transform due to its property of retaining the details present in the data.
In DWT, the most prominent information in the signal appears in high amplitudes and the less prominent information appears in very low amplitudes. Data compression can be achieved by discarding these low amplitudes. The wavelet transforms enables high compression ratios with good quality of reconstruction. At present, the application of wavelets for image compression is one the hottest areas of research. Recently, the Wavelet Transforms have been chosen for the JPEG 2000 compression standard.
. FEATURE EXTRACTION
A feature is a characteristic that can capture a certain visual property of an image either globally for the whole image, or locally for objects or regions. Content based image retrieval, a sub domain of computer vision, is a system in which a computer analysis an image to extract visual features. These features are known as low level features. Some key issues related to CBIR systems are the following, first how the extracted features can present image contents. Second,
how to determine the similarity between images based on their extracted features. One technique
for these issues is using vector model. This model represents an image as a vector of features and
the difference between two images is measured via the distance between their feature vectors.
Feature extraction module extract and save image features to the feature database automatically. Texture is one of the most important features for CBIR. Texture refers to the visual patterns that have properties of homogeneity not resulting from presence of only one color or intensity. Texture features are extracted from co-occurrence matrices and wavelet transforms coefficients.
This paper has shown how one can use new transform is complex wavelet transform (DST) to enhance the image retrieval process. They have shown that we can achieve almost the same precision for color image retrieval as well. These properties of CWT have motivated us to use it as feature extraction for our proposed system.
MEDICAL IMAGE DATABASE
A CT scan shows detailed images of any part of the body, including the bones, muscles, fat, and organs. Spatial and contrast resolution are dependent on the energy of the x-ray source, slice thickness, field of view, and scanning matrix. High resolution CT provides excellent delineation of osseous structures.
In this system six different categories of CT scan images used for retrieval, 20 images in each category so total 120 images store in database from that one image of each group shown in figure (5.1). This data collect from Nobel hospital, Pune and some of the images available at internet. Each image has different size but we can convert in fixed size form by using Matlab command resize that is 256 X 256 size.
CONCLUSION
In this paper it was proposed to extract features using Discrete Sine Transform (DST) and select the top 50 attributes based on class attribute using information gain. The extracted
features were trained and classified with SVM using poly kernel. A novel SVM was proposed and the classification accuracy of the proposed method improves by a factor of 5.18.
The reduced features in the proposed method, decreases the overall processing time for a given query input image.