Pixel Level Weighted Averaging Technique for Enhanced Image Fusion in Mammography

Pixel Level Weighted Averaging Technique for Enhanced Image Fusion in Mammography Abstract M Prema Kumar, Associate Professor, Dept. of ECE, SVECW (A), Bhimavaram, Andhra Pradesh. P Rajesh Kumar, Professor & HOD, Dept. of ECE, AUCE (A), Andhra University, Visakhapatnam. premkumarmedapati@gmail.com Image fusion at the basic level is combining two images to obtain a new image with enhanced features. Mammography is a process of imaging technique used to identify breast cancer. In this paper a simple pixel level averaging technique is used for image fusion. A single X-ray is taken and enhanced to produce a new image. The enhanced image is fused with original image using simple pixel averaging technique. The observed results are quite better when compared with the single X-ray. Keywords:Mammography, pixel level image fusion, image fusion Introduction Image fusion is the process of combining multiple images of a scene into a single composite image that contains all the important features from each of the input images. The resultant fused image can provide a more accurate and reliable information about a scene than any of the individual source images (Blum & Liu 005). The fused images obtained are more suitable for human visual perception and to carry out subsequent image processing tasks such as segmentation, classification, object detection, tracking and identification (Cvejic et al 007, Liang et al 013). Recently, image fusion is extensively useful in applications such as surveillance, remote sensing, military, machine vision, robotics and medical imaging.the fusion of redundant and complementary multiple sensor information reduces the overall uncertainty and increases the accuracy. The use of multiple sensors result in large amount of data and image fusion reduces the amount of dat thereby reducing the storage requirements. The important advantages of image fusion are that the image information can be obtained more accurately, as well as in less time and at a lower cost (Blum & Liu 005). In general, a fusion scheme should satisfy the following requirements (Rockinger 1996, Piella 003): It should identify the most significant features in the source images and transfer them without loss of detail into the fused image. The image fusion process should not introduce any artifacts or inconsistencies which can mislead or divert a human observer or further processing tasks. It should be reliable, robust and suppress the irrelevant parts of the image and noise. Medical image fusion is the technology that could compound two mutual images in to one according to certain rules to achieve clear visual effect. By observing medical fusion image, doctor could easily confirm the position of illness. Medical imaging provides a variety of modes of image information for clinical diagnosis, such as CT, X- ray, DSA, MRI, PET, SPECT etc. Different medical images have different characteristics, which can provide structural information of different organs. For example, CT (Computed tomography) and MRI (Magnetic resonance image) with high spatial resolution can provide anatomical structure information of organs. And PET (Positive electron tomography) and SPECT (Emission computed tomography) with relatively poor spatial resolution, but provides information on organ metabolism [3] [6]. Thus, a variety of imaging for the same organ, they are contradictory, but complementary and interconnected. Therefore the appropriate image fusion of different features becomes urgent requirement for clinical diagnosis. In this paper a simple pixel averaging technique is used to fuse images to produce an enhanced. The radiologist will be able to diagnose breast cancer more easily as the fused X- ray will have more information when compared to a simple X-ray. In general the process of mammography has some stress to be undergone by the female subject. And when the process of image fusion is considered, it has to have two or more s from the same subject for fusion process. To reduce the stress a new enhanced fusion method from one is proposed in this paper. Pixel Level Image Fusion Pixel level fusion can be used to increase the information content associated with each pixel in IJIEE 015 10

an image formed through a combination of multiple images, e.g., the fusion of a range image with a two-dimensional intensity image adds depth information to each pixel in the intensity image that can be useful in the subsequent processing of the image. Different images to be fused can come from a single imaging sensor or a group of sensors. The fused image can be created either through the pixelby-pixel fusion or through the fusion of associated local neighborhoods of pixels in each of the images. The improvement in quality associated with pixel-level fusion can most easily be assessed through the improvements noted in the performance of image processing tasks such as (segmentation, feature extraction, and restoration) when the fused image is used to compare the use of the individual images. The fusion of multisensory data at the pixel level can serve to increase the useful information content of an image so that more reliable segmentation can take place and more discriminating features can be extracted for further processing.pixel-level fusion can take place at various levels of representation: the fusion of the raw signals from multiple sensors prior to their association with a specific pixel, the fusion of corresponding pixels in multiple registered images to from a composite or fused image, and the use of corresponding pixels or local groups of pixels in multiple registered images for segmentation and pixel-level feature extraction. Fusion at the pixel level is useful in terms of totalsystem processing requirements because fusion is made of the multisensory data prior to processing-intensive functions like feature matching, and can serve to increase overall performance in tasks like object recognition because the presence of certain substructures like edges in an image from one sensor usually indicates their presence in an image from another sufficiently similar sensor. In order for pixel-level fusion to be feasible, the data provided by each sensor must be able to be registered at the pixel level and in most cases, must be sufficiently similar in terms of its resolution and information content. Although it is possible to use many of the general multisensory fusion methods for pixel level image fusion. In this paper a simple pixel level weighted averaging technique is used for image fusion of mammography images. A comparison of simple average and weighted average is performed using image quality assessment techniques. The fusion of images is performed using weighted averaging method as per the following equations [5]: Pixel Level Average method based image fusion Image averaging is the most commonly used arithmetic pixel based fusion method in which the fused image is obtained by the pixel-by-pixel averaging of the input images. I( = ---------------- (1) Pixel Level Weighted Average based image fusion Weighted averaging image fusion methods generate a fused image pixel by pixel, as an arithmetic combination of the corresponding pixels in the source images as given by Equation W I( = where I(! Fused image I 1 (! Input image 1 I (! Input image W 1, W! Weights added -------() The end fused image is measured for quality with the Image quality measureing parameters Image Quality Parameters The image quality parameters as listed are used for analyzing the fused images [6] Signal to Noise Ratio (SNR), Peak Signal to Noise Ratio (PSNR), Root Mean Square Error (RMSE), Normalized Cross Correlation (NCC), Mutual Information (MI) Universal Image Quality Index (UIQI), Fusion Factor (FF), and Fusion Symmetry (FS) The parameters are tabulated in Table1 and Table for the methods proposed in this paper. Results I1( + I ( 1 * I1( + W * I ( W + W The method is tested for three types of X-ray s namely, Normal, Benign and Microcalcification X-ray s. The output images for different inputs using simple average method and weighted average method are shown in figures, 3, 4, 5, 6 and 7. The same were also assesed using the image quality parameters and tabulated as in Table 1 and Table. IJIEE 015 11 1

Table 1: Image fusion using Pixellevel averaging method, parameter analysis S.No. Parameter Normal X-ray X-ray Mammogram with Microcalcification 1 RMSE 55.93 44.6618 6.8591 PSNR 13.1791 15.130 19.5489 3 NCC 0.9951 0.6559 0.983 4 MI.1654.6577 1.9133 5 IQI 0.8198 0.5061 0.9454 6 FS 0.0188 0.0871 0.0348 Table : Image fusion using Pixel Level weighted averaging method, parameter analysis S.No. Parameter Normal X-ray X-ray Mammogram with Microcalcification 1 RMSE 39.974 35.639 5.1409 PSNR 16.0955 17.0937 0.13 3 NCC 0.457 0.698 0.7904 4 MI 0.856 1.351 1.9385 5 IQI 0.5405 0.511 0.8190 6 FS 0.0708 0.0516 0.0354 Figure 1: Image fusion using Simple Average method for Normal Breast Mammogram (a) Normal Breast X-ray image (e) Fused image of normal and histogram equalized image (c) Histogram of Normal Breast IJIEE 015 1

Figure : Image fusion using Simple Average method for Benign Breast Mammogram (a) image (e) Fused image of Benign and histogram equalized image (c) Histogram of Benign Breast Figure 3: Image fusion using Simple Average method for Microcalcification Breast Mammogram (a) Microcalcification Breast X- ray image (c) Histogram of Microcalcification Breast Figure 4: Image fusion using Simple Weighted Averagemethod for Normal Breast Mammogram (e) Fused image of Microcalcification and histogram equalized image (a) Normal Breast X-ray image (e) Fused image of normal and histogram equalized image (c) Histogram of Normal Breast IJIEE 015 13

Figure 5: Image fusion using Simple Weighted Averagemethod for Benign Breast Mammogram (a) image (e) Fused image of Benign and histogram equalized image (c) Histogram of Benign Breast Figure 6: Image fusion using Simple Weighted Averagemethod for Microcalcification Breast Mammogram (a) Microcalcification Breast X- ray image (e) Fused image of Microcalcification and histogram equalized image (c) Histogram of Microcalcification Breast Pixel level average method 60 50 40 30 0 10 0 RMSE PSNR NCC MI IQI FS Normal X-ray X-ray Mammogram with MicrocalcificaFon Figure 7: Parameter Analysis of Pixel level simple averaging method IJIEE 015 14

50 Pixel level weighted average method 40 30 0 10 0 RMSE PSNR NCC MI IQI FS Normal X-ray X-ray Mammogram with MicrocalcificaFon Figure 8: Parameter analysis of pixel level weighted average method Conclusion: The two methods of simple averaging and weighted averaging technques have been tested on three types of s i.e., Normal, Benign and Microcalcification s collected from MIAS database. It was observed that a significant change of results was observed in the image quality parameters. It is concluded that by using these image fusion methods, the enhanced images will enable the radiologist in diagnoising the breast cancer in a better and easier way. References: [1] Jagalingam P., ArkalVittalHegde, Pixel Level Image Fusion A Review on Various Techniques, 3rd World Conference on Applied Sciences, Engineering & Technology 7-9 September 014, Kathmandu, Nepal. [] A. P. James, B. V. Dasarathy, Medical Image Fusion: A survey of the state of the art, Information Fusion, 014 [3] Z. Wang,C.A. Clavijo, E. Roessl, U. van Stevendaal, T. Koehler, N. Hausergy and M. Stampanoni, Image fusion scheme for differential phase contrast mammography, 7 th Medical Applications Of Synchrotron Radiation Workshop (MASR 01) Shanghai Synchrotron Radiation Facility (SSRF), 17 0 October, 01, Published By IOP Publishing For Sissa Medialab. [4] V.P.S. Naidu and J.R. Rao, Pixel level image fusion using wavelets and principal component analysis, Defence Science Journal, Vol. 58, No. 3, May 008, pp. 338-35. [5] V. Jyothi, B. Rajesh Kumar, P.K. Rao, D.V.R.K. Reddy, Image Fusion using Evolutionary Algorithms (GA), International Journal of Computer Technologies and Applications, (), 01.. [6] M Prema Kumar and P Rajesh Kumar, Image Fusion of Mammography Images using Genetic Algorithm (GA), Australian Journal of Basic and Applied Sciences, 9(33) October 015, Pages: 45-50. IJIEE 015 15