In current scenario back pain is a leading

cause of work absenteeism. Another area of concerns is the cervical spine

trauma which causes majority of spinal lesions. Many of these problems are the

consequence of an abnormal spinal motion. In past a Greek method was applied for

placing the displaced bones at their original positions. This method was known

as automatic method as no machines were used and force was applied manually.

Then with the advancement automatic traction therapy was replaced by mechanical

traction. The Skin traction was used at a great extent during the Civil War for

fractured femurs (thigh bone). It was commonly known as the “American Method

for treating bones fractures and dislocation”.

Now a day’s Decompression Traction Therapy

becomes an effective treatment for spinal traction treatment. It is less

expensive than surgery. Decompression Traction Therapy is the new therapeutic

device for treatment of painful nerve compression and disc herniation

syndromes.”

Mainly we can perform traction either in

continuous manner or an intermittent manner. Usually static traction is used

for cervical spondylosis and osteoarthritis or degenerative disc disease DDD

and intermittent traction is common for Facet joint Dysfunction and joint hypo

mobility. Both of the intermittent and the continuous cervical traction had a

significant effect on neck and arm pain reduction, a significant improvement in

nerve function and a significant increase in neck mobility 1

Cervical traction is a method in which a distracting force is administered to

the neck so as to separate the cervical segments and relieve compression of

nerve roots by intervertebral disk distraction 2. Traction may

improve conduction by improving blood flow to cervical nerve roots. Cervical

traction has a significant biomechanical effect on spinal structures, which can

be demonstrated by CT evaluation before and after traction. 3 One practical device

to extract spine motion in real-time is the digital video fluoroscopy (DVF). 4

Advances in

medical image processing have led to increase in large image collections.

Hence, techniques that are computer assisted are becoming more promising

approach in medical image analysis. The feasibility of computer assisted

techniques for the segmentation of vertebral bodies in spine X-ray images has

been of great interest 24, 25 to biomedical researchers. Reliable extraction

of vertebrae boundaries is a prerequisite for subsequent pathology validation

and Content-Based Image Retrieval (CBIR) research. However, fully automated

segmentation of spine X-ray images is a very challenging problem.

A smart approach is discussed for cervical

traction which is a combination of image Processing and Physics theory. A

fundamental issue faced in the design of image analysis techniques is the

identification and characterization of the image space. In plain words, one

ends up having first to answer the question “What do we mean

mathematically by images?” A pertinent example of this point can be found

in the variational approach to image de-noising, the goal of which is to

estimate the original version of an image from a given degraded one. The

variation approach to this problem seeks to exhibit the “restored”

image as the minimize of a functional defined over the space of all images. The

first task is clearly to decide which space of functions to take images from.

For instance, it is easy to see why Sobolev spaces are ill suited for this

purpose: their elements cannot have discontinuities across co-dimension one

surfaces, and yet any successful model of images should allow for them. Such

discontinuities need to be allowed because one of the most important features

of images, namely “edges” (places where one object in the scene ends

and another begins) correspond squarely to this type of behavior. The space of

functions of bounded variation therefore provides the more appropriate setting.

(Another example comes from the variational approach to the image segmentation

problem, where the correct space of functions for minimizing one of the most

successful models in this context, the Mumford-Shah segmentation model, turns

out to be a subset of functions of bounded variation, known as special

functions of bounded variation) 32. Roughly the various approaches to image analysis is divided into

three categories: (A) Statistical representations (B) Spectral and wavelet representations,

and (C) Scale-space representations.

Statistical approaches treat images as

samples from random fields, which are often modeled by Markov/Gibbs fields or

via statistical learning from an image database. This approach was pioneered in

the 80’s by Grenander (Brown) and the Gemans (Brown and John Hopkins.

Spectral and wavelet representations are the

mathematical foundation for JPEG Internet image coding protocols. It has uses

beyond image compression. In fields like computer vision and medical imaging,

there is great demand for efficient and accurate image processors.

Transform

Used

Making of

Model

Fig: Designing of Image Processor

Major

task is how to design an image processor that performs efficiently and well?

Typical tasks would be: denoising, edge detection, intensity enhancement, and

compression and decompression. In addition to these relatively low-level tasks,

there are mid- and high-level tasks like image segmentation, and pattern

identification and recognition.

Initially

a suitable model is constructed for the given task. Currently, the developed processing models

(Fig) are using tools like Bayesian 34, 35 decision, inverse problems,

variational optimization, etc. The need of these approaches arise mainly in the

fields of statistical mechanics, nonlinear

PDEs, differential geometry and topology, the calculus of variations, ,

harmonic and functional analysis, and numerical analysis. Image denoising is

one of the major problem faced during analysis of image. Numerous methods has

been proposed for image denoising problem, from using transformations and

statistical methods to using PDEs. For transformations, typically spectral transformation

are used and recently Curvlet or Ridgelet transformation are used. Another problem

in Image reconstruction is inpainting. The Digital inpainting problem, is

related to disocclusion problem. There are various approaches for the image

segmentation problem: using snake methods or region growing and merging

techniques or global optimization approaches using energy functional or

Bayesian approaches 33, 34.

When one

has a processing model, the next stage is its analysis, keeping in mind the end

goal to answer inquiries like presence and uniqueness, dependability, properties

of solutions etc. Many image processing model are nonlinear and they requires

new mathematical insights, they require rigorous mathematical analysis. As

there are many transformation algorithms present ,the major question arises is

which among the many techniques proposed for the same task is superior, what

effect the various parameters that appear in a method have on its behavior, and

under what conditions a given technique can be expected to perform well.

Finally,

what is an effective algorithm for implementing the image analysis model? One

can certainly create, implement and plan schemes without performing the type of

numerical analyses described above. But one runs the risk of developing an

algorithm that performs ineffectively in the presence of noise. Also, the

history of scientific computing shows that the breakthroughs that led to

massive speed-up would have been unimaginable without a profound understanding

of the mathematics. Mostly the mathematical calculations can be converted into

efficient and robust algorithms. The methods currently being used in image

processing come from almost all branches of scientific computing including fast

fourier and wavelet transforms, multigrid algorithm, dynamic programming,

combinatorial optimization, computational PDEs, NLA, and Monte-Carlo

simulations.

We can

use the important step of image preprocessing i.e image segmentation to analyze

how mathematics has contributed to image processing in the stages discussed

above: modeling, analysis, and implementation 36. The goal of segmentation is

to divide up the image domain into as smaller segments as possible, so that

image features (such as gray-scale intensity, or color) either slowly varying

in each piece or approximately constant. This procedure helps differentiate

between parts of the image domain occupied by distinct objects in the scene. It

is a challenging problem that is intimately connected with edge detection.

Typically, physics plays a vital role in

cultivating the knowledge of natural sciences and establishing the foundation of logistic and scientific

methods. The application of physics to medical sciences has been developed in

practice for centuries 5.

X-Ray, CT scan and MRI images of the spine provide a practical approach for

detecting and assessing vertebral abnormalities. For accurate vertebra

detection number of features of an image can be extracted and meaningful

information can be obtain from low level information in the image using

segmentation. Image segmentation is observed to be the maximum important part

in digital image processing. Segmentation is nothing but however a portion of

any photo and object is extracted and analyzed to get useful information. In

image segmentation, digital photo is splitted into a couple of set of pixels. Main objective of segmentation is to

partition the image into the meaningful areas 6.

The segmentation techniques used for medical images are specific to the type

body part and imaging modality to be used. Consistent algorithms are necessary for the explanation of

anatomical structures and other regions of interest (ROI).As the time grow

algorithmic complexity increases. As medical

images are complex and rarely have linear feature, segmentation of medical

images becomes a difficult task. There is no universal algorithm which can give

appropriate results for every image. While using segmentation algorithm

different artifacts and noise introduce by electronic system must be taken in

consideration 7.

A comparison of three generation medical image segmentation technique is

provided in literature 8.

Vertebrae localization is a difficult task

to perform manually. Model-based approach is used for vertebra detection,

identification and segmentation in the case of CT images 9.

Each cervical vertebra has a unique morphology, which make quantitative

analysis a challenging task. So developing an approach which is robust,

efficient and capable of robustly determining the location of the spine and the

contour of the vertebrae is essential. Author proposes fully automatic

framework for cervical spine mobility analysis on radiographs 10 which is based on

-fully automatic vertebra detection, vertebra segmentation and vertebra

mobility analysis.

Digital medical images have some degree of

noise it gets corrupted with noise during its acquisition from various sensors

and transmission process. Since

different types of images have different characteristics and different

applications have different requirements, the filtering algorithms should be

designed for each case properly. Mostly the medical images quality cannot

satisfy the medical analysis and the applied requirements. The denoising of

medical image plays an important role in the image processing, and it is the

basic of further analysis and computation. The image edge are easily

interrupted by noise, but the traditional image denoising smooth out the edges

of the reconstructed images which caused edge to be blurred, and made the

information lost. The

denoising algorithms based on partial differential of total variation is

proposed 11. Author ensures that

the algorithm works to obtain clear medical image and preserve the edge

information integrally at the same time.

Mostly, the

medical images quality cannot satisfy the medical analysis so to get best

quality image a set of textures and images is analyzed to determine the best

measure of image activity and it has showed that image activity measure has

powerful ability to capture the activities and differentiating between various

images 12.

To detect degenerative changes which occurs due to narrowing of intervertebral disc space author introduce a

machine learning approach to detect and localize degenerative changes in X-ray

image analysis of cervical spine and author receive accuracy of 95%for a patient

data set of 103patients 13. Another author

discussed the same problem of detecting degenerative changes in particularly

C3-C6 vertebra of cervical. The result range from 82% to 94% ,author proposed segmentation

and clustering from where distance features are computed and mapped 14.

To preserve edges and fine details while

effectively removing noise, both local gradient and variance are incorporated

into the diffusion model 15 16. A newer

technique is introduced for noise removal using fractional-order anisotropic

diffusion equations, which uses the discrete Fourier transform and an iterative

scheme in the frequency domain 17.

Image filtering is an effective means for improving image quality. Author

proposed local activity-tuned filtering frameworks for noise removal and image smoothing,

where the local activity measurement is given by the clipped and normalized

local variance or standard deviation. Local activity-tuned relative total

variation framework achieves good performance for image smoothing and

represents the image in different scale-space and it has been used for depth

image denoising 18.