Visualization of Hidden Structure and Shape in Ct Image via Non-Linear Perspective Foreground and Back Ground Projection

:- The neurologist analyse the brain images to diagnose the disease via structure and shape of the part in the scanned Medical images such as CT, MRI, and PET.The Medical image segmentation perform less in the regions where no or little contrast,artefacts over the different boundary regions. The manual process of segmentation show poor boundary differentiation dueto discernibility in shape and location, intra and inter observer reliability. In this paper, we propose a dyadic Cat optimization (DCO) algorithm to segment the regions in the brain from CT and MRI image via Non-linear perspective Foreground and Back Ground projection. The DCO algorithm remove the artefacts in the boundary regions and provide the exact structure and shape of the brain regions. The DCO algorithm show the region boundary such as plerygomaxillary fissure, occipital lobe, vaginal process zygomatic arch, maxilla and piriform aperture with high visibility in the regions of inadequately visible boundary and distinguish the deformable shape. The DCO algorithm show the increased SSIM and 90 percent accuracy.


Introduction
In medical imaging, CT image prone to artifact and low contrast in the boundary region due to reconstruction of image from independent detector located in order for measurement. The artifact classify into four types such as physics based artfifacts,patient-based artifacts, scanner-based artifacts and helical and multisection artifacts.The physics based artifacts arise during the acquisition of data from CT. The patient based artifacts arise due to movement of the patient during scanning. The scanner based artifacts cause due to improper function of scanner in CT machine. In addition, streaks appear in CT image because of dense object. The dense objects in the image lack in contrast and clear boundary, when X-ray beam pass through the object the photon with lower energy absorb more quickly than Photon with higher energy. The brain structure change along with the person age.The changes in structure various with individuals and the person with same age.Furthermore, for detection of neurodegenerative disease, such as Alzheimer's disease (AD),Parkinson's disease (PD),Prion disease,Motor neurone diseases (MND),Huntington's disease (HD),Spinocerebellar ataxia (SCA) and Spinal muscular atrophy (SMA) need the clear boundary edge and shape of the brain organs to identify the deformation. The neurodegenerative disease easily identify with the brain structure variability from CT atlas via nonparametric and parametric approach.
The structure in brain measure through the CT or MRI imaging with some correction or adjustment in size variation. Such variation obtain through ventricle-brain ratio, in which the size of the brain structure is proportion to the head size estimation. Furthermore, regression analysis and volumetric measure technique apply for structure measures, which come under the parametric measurement.However,the non-parametric measure execute through the ROI, segmentation algorithms which perform both manually and automatically. The brain structure measurement through parametric and non-parametric algorithm in MRI and CT image has certain disadvantages such as resolution ,beam hardening artifacts, view limit in posterior fossa and low visualization in white matter disease for CT image and the disadvantages of MRI for structure of brain such as Claustrophobia, long exam and scanning is ineligible for the patients with pacemakers. Both CT and MRI provides the physical structure of the brain in the static image. However, CT image provide the detail information about spinal cord or brain in which x-rays passing through the organs in different angles. The x-rays are attenuated in the regions of high density materials such as bone and calcium, where the structure of these regions appear white and the structure which permit the x-rays looks in the dark appearance in the image. The attenuation, beam Harding , artifacts, steaks lead to inaccurate measurement of structure in the brain, due to unclear boundary region, shape and edges in the CT image.
In this paper, the proposed DCO method contribute for the superior neurodegenerative diagnosis via brain structure analysis and measurement from CT image.
(i) Improve the manual and automatic segmentation in the regions with little or no contrast in the boundary region due to artifacts and steaks (ii) The hidden such sub structures such as thalamus and white mater are fully discernible through the Non-Linear Perspective Foreground and Back Ground Projection.
(iii) Provides the spatial relationship between the structures in the brain and identifies the deformation of the structure via specific boundary condition such as length and shape.

RELATED WORK
The segmentation is especially well suited for structures with weakly visible boundaries as it simultaneously estimates the image inhomogeneities, explicitly models the boundaries through a deformable shape model, and segments the MR images into anatomical structure [1]. Furthermore, software platform integrates algorithms for image analysis and processing. The image-processing block includes segmentation, visualization, reconstruction and registration for image analysis.However, Conventional Quality assessment of 3D image segmentation algorithm done by comparing segmentation quality with datasets. The comparison determines systematic segmentation problem. The authors propose a clustering algorithm to compare quality of segmentation with different 3D images [2]. Furthermore, connectivity criterion applies with clustering algorithm for better segmentation quality assessment [3]. Furthermore, Organ medical image segmentation requires prior knowledge of location and appearance of organs. The detailed prior knowledge of organ shape is challenging. The solution space of organ in 2-D image determine with MAP-MRF contour segmentation method. The contour has prior knowledge of boundary -edges, shape and appearance [4]. Furthermore, a master slave decomposition help in labelling each iteration. However, conformal geometric algebra (CGA) provide optimal solution for complex geometric problems. The conformal geometric algebra, highly inconvenient due to computational complexity and high dimensionality [5]. expert's decision on image data. The segmentation results acquire better by closed loop system comprising of automated boundary detection with human interface for better judgement [10]. Furthermore, Bias field apply to correct intensity in MRI images. A modified version of Mumford-shah performs segmentation and bias correction simultaneously. Initially the L0 gradient regularizes to model bias field. The segmentation step comprises of two steps. The first step recovers intensity of bias field. The second step applies thresholding for image segmentation [11]. Medical image comprises of complex foreground and background density distributions. Noise and low contrast further make it hard for boundary identification in foreground and background images. A SVLS (Supervised variation level set) apply to differentiate intensity differences in foreground and background images [12]. Segmentation of 3D High-frequency Ultrasound Images of Human Lymph Nodes Using Graph Cut with Energy Functional Adapted to Local IntensityDistribution. A 3D quantitative ultrasound imaging differentiates cancer free lymph nodes from metastatic lymph nodes. The methodperform by automatic segmentation method to differentiate fat from ultrasound processing for attenuation. The lymph nodes and the fat have varying intensity [13].

NON-LINEAR PERSPECTIVE FOREGROUND AND BACK GROUND PROJECTION
Dyadic wavelet transform form with low pass, high pass filters and down-sampler. The dyadic wavelet transform obtain by convolution of dyadic dialects with mother wavelet against original image. Due to zero mean and finite support nature of mother wavelet the shape equivalent to an edge.
The dyadic transform dilation represent by The wavelet transform at a particular location x define by

W W
In addition, the filter bank in dyadic wavelet transform produces orthogonal coefficient from even output of high pass filter and coarse coefficient even output of low pass filter. The coarse coefficient expands by dilation to form 2D discrete Dyadic wavelet transform. The different dilation factors perform non-linear perspective foreground and back ground projection. Boundary ,edge detection at different levels in the image. The major edge detection identifies at all dilation levels and remove darkness in the boundary region due to artifacts and steaks. The minor edge detection identifies only at low dilation levels. For uniform identification of edges at all levels in M dimensional space CAT optimization apply.

EDGES AND BOUNDARY OPTIMIZATION:
The CAT optimization algorithm operation is twofold namely seeking mode and tracing mode. In seeing mode, the CAT rests looking for change in edges. The change in edges represents by change in count of dimension (CDC), seeking memory pool (SMP), self-position consideration (SPC) and selected dimension (SD). If change in edges detect the CAT change to tracing mode. In tracing mode, the CAT position change with respect to change in position, speed and direction of edge.
In seeking mode,the SMP make copies of CAT and if SPC is true the position of CAT is saved.  Cat k is in the seeking mode?
Apply cat k into seeking mode process Apply cat k into tracing mode process Re-pick number of cats and set them into tracing mode according to MR, and set the others into seeking mode. Terminate?

End Yes No
No

Yes
The dyadic wavelet transform followed by CAT swarm optimization algorithm apply on CT medical image. The Figure 3 shows input CT image. CT image shows bone, muscle and organ masses. The CT image represent in grayscale image.   1500. This is due to the fore ground image pixel weight is more.

Conclusion
Neurologist asses the brain images to diagnose the disease via structure and shape of the part in the scanned Medical images such as CT, and MRI. The MRI and CT image apply with novel DCO algorithm for the anatomical structure and shape in perspective projection in non-linear approach to show the foreground and Background of the regions. The algorithm suit for both MRI and CT image for the region where the boundaries are less visible and to interprets the deformation shape of the anatomical structures in the brain.The anatomical structure of the brain image show the improvement in the shape and boundaries due to the dilation property of the DCO algorithm and exact edge due to the count of dimensions. Furthermore, the algorithms can be applied in the PET and ultrasound image for the week boundary regions for precise shape and deformation detection.

Funding
: No funding.