Super Resolution Image Reconstruction for Bone Images

Objectives: Images got from image intensifier to assess joint movements during surgery are usually distorted. This paper aims to get exact high resolution image from such low resolution (distorted, blurred) images using super resolution image reconstruction. Methods/Statistical analysis: This method uses different images taken from the video of the original procedure of studying a patient with stiff hip to assess his hip movement under image intensifier. The proposed method uses the Singular Value Decomposition (SVD) transformation model for the purpose of reconstruction. In this work two data sets of images extracted as video from a c-arm –image intensifier imaging of a patient with disease of left hip are analysed. Findings: While capturing a fast moving object needs a camera with more radiation. Image intensifiers-even with methods like gating have disadvantage of smear phenomenon when capturing fast moving objects. But clear, quick analysis is needed in operation room setting where time is vital. In this method details like the inferior cortex of the neck of femur or proximal end of the trochanter which is absent in captured figures are appreciated in output high resolution image. Thus the outputs generated without omitting essential details. Application/Improvements: Distorted images captured during real-time procedures in the operation theatre while patient’s part is moving is resolved using Super Resolution Image Reconstruction technique. Thus surgeon understand the patient’s bone better. Procedures like placing a screw or passing a guide wire to reduce a fracture can be easily done with such a methodology. *Author for correspondence


Introduction
Imaging is an integral part of management of patients. It is used in three areas viz 1. Before the orthopaedic procedure to select a case, 2. During the procedure to properly guide the surgeon to do the procedure well and 3. After the procedure to assess the effectiveness of the procedure   and assess the results of the procedure. Even as the surgery is on, during the procedure movements of the limbs during the procedure happens and sometimes it is necessary. So also passage of the implants within the patient's body also involves movements. Capturing such moving objects and producing them as images is routinely done by machines like image intensifiers 1 .
Keywords: Bone Images, Image Intensifier, Image Reconstruction, Moving, Super Resolution While assessing the images of a moving limb in the operation theatre the surgeon is left with a set of video images which are blurred and fast in progression. The judgement of the doctor is difficult for example as to what is the limit of the actual bone on which the operation is done and if the position of the implant inside the bone is correct.
The commercially available cameras have steeper gradation which is suitable for fast moving objects. This is useful for faster moving objects like cardiac angiogram etc. the television images are having rich contrast and are free form after glow. However, the disadvantages are that in the imaging intensifier there can be noise high gamma value -high contrast and hence more radiation exposure to the patient and the team which is operating. Price of such image intensifiers are also high e.g. the plumb icon tube. While another device the vidicon is effective in capturing non-moving and slow moving object. Also vidicon is cheap with less radiation 1 . Thus to capture a fast moving object one needs a camera with more radiation. Image intensifiers are having the disadvantage of the smear phenomenon when capturing the objects which are moving fast 2 . Gating is a method employed in the image intensifier to reduce the motion blurr 3 . Even after such employing such methods still there are blur in the images which needed to analyse in the operation room setting, where time is important. It is in this background, we hypotheses that if images of good resolution are available to the surgeon even with moving limbs captured with the same image intensifier, then the result of the surgery will be even better. To our knowledge, there is no such article which discusses the resolution using moving limb recorded as video.
Super resolution image reconstruction is a concept by which images with poor clarity can be used to provide images of good quality for the end user.
In this work we have applied an algorithm to simplify the work of the surgeon in understanding the image captured during the movement of the limb during a procedure.

Methodology
A patient with suspected disease of left hip is taken for the study. In this disease, the head of femur bone (thigh bone) will be deformed and hence a mechanical block of the hip movements is expected. To check the range of movements objectively, a c-arm -image intensifier is routinely used. During this procedure, for comparison, the normal side hip was also imaged and its videos were recorded for dynamic assessment of the hip movements of normal side. The videos were in mp4 format. The current study is done on the images taken from these videos recorded using Surgico Hfx of Medtronics -Philips. In

Enhanced Singular Value Decomposition Fusion
The   (2) This is done prior to the interpolation step because interpolation may cause some loss in the information.

Analysis of the Images
The output image has incorporated the details in both first and second input data. The problem of visualising the moving bone while doing a real time procedure is common. This method uses different images taken from the video of the original procedure and the outputs generated without omitting essential details. Procedures like placing a screw or passing a guide wire to reduce a fracture can be easily done with such a methodology.
Consider the Figures 3, 4, 5 and 6; the inferior cortex of the neck of femur (marked with red arrow) which is absent in Figure 4 but present in Figure 3 is appreciated in Figure 6 the output high resolution image.
Also the proximal end of the trochanter (marked with blue arrow) which is prominently seen in Figure 4 is not prominently seen in Figure 3 is appreciated in Figure 6 the output high resolution image.

Conclusion
Images even when captured while patient's part is moving can be resolved using Super Resolution Image Reconstruction technique. Clear and exact images thus obtained give good idea on the patient's bone to the surgeon during real-time procedures. This will give the suffering patient a better care.