Research On Multiscale Entropy Analysis For Infrared Image Of Human Lumbar Disease

In recent years, with the development of infrared thermal imaging technology, itsapplication is more widely in the medical field. However, using the infrared thermalimager to human lumbar disease research is not in-depth. In the paper, we use infraredcameras G100to shoot the human lumbar spine. Through a series of preprocessing,namely selecting the image, reducing image noise and improving image quality, theinfrared image is converted into the pixel sequence. Then, we use multiscale entropymethod to analyze the sequence of pixel values and by the entropy curve to show thepresence or absence of disease and the extent of the lesion. From the infrared imageacquisition experiments, infrared image preprocessing and multiscale entropy method forinfrared image processing three aspects build human lumbar detection software. It canprovide a reference for the clinical diagnosis. The main research of this paper is describedas followed.1) Human lumbar infrared image capture experimentsThere are three experimental parameters need to be setted, they are environmentaltemperature, humidity and parameter selection infrared camera.With the help performancecorrection of thermal infrared imager, the infrared image will achieve the best results.2) Analysis of infrared thermal imageFirstly, Selection of infrared image make lumbar area is full of the whole image;Secondly, median filtering and the weighted histogram equalization are used to reduce thenoise and enhance the effective information of the image. Finally, we use the multi-scaleentropy algorithm based on sample entropy analysis to process the pixel sequence ofinfrared image. This topic verifies the feasibility of this method and finds the upper andlower values of the normal entropy curve about normal lumbar infrared thermal images. Inmost of the scales, if lumbar infrared thermal image entropy value is less than normalentropy curve, it shows that the lumbar area have lesions, And the smaller the entropyvalue is, the more serious the lesions.3) By using of MATLAB image processing tool and GUI function, human lumbardetection application software is designed.