Hisashi Shimizu

An automatic detection method of spinal deformity from moire topographic images employing asymmetric degree of shoulder and waistline

Spinal deformity is one of serious diseases, mainly suffered by teenagers during their growth stage. To detect the spinal deformity in early stage, orthopedists have traditionally performed a painless examination called a forward bending test or moire topographic image test in mass screening of school. It is, however, inspection base on the forward bending method takes much time and moire images, and visual examination also require a large amount of moire images because they are collected from elementary as well as junior high schools. This causes exhaustion of doctors and therefore leads to misjudgment. Therefore realization of automatic spinal deformity detection based on the moire images has long been desired among orthopedists. In this paper, we propose a method for automatic detection of spinal deformity from moire topographic images by using a new asymmetric feature which is obtained by using statistical features between left- and right hand side shoulder and waist lines. We classified an unknown moire image employing linear discriminant function. The proposed technique is applied to 1200 real moire topographic images. By the employment of the two asymmetric features, 66 % of unknown moire images were successfully classified in the performed experiment. Some experimental results are shown along with discussions.

Automatic spinal deformity detection based on asymmetric degree of moire topographic images

Spinal deformity is a disease mainly suffered by teenagers during their growth stage particularly from fifth year in the elementary school to second year in the middle school. For checking the spinal deformity detection, a large amount of moire images were collected from elementary as well as junior high schools by orthopedists. This causes exhaustion of doctors and therefore led to misjudgment. Therefore realization of automated spinal deformity inspection based on the moire images has long been desired among orthopedists. In this paper, we propose a technique for spinal deformity detection based on difference of symmetric degree on moire topographic images. The technique can classify as normal or abnormal from the inputted moire topographic images automatically.

Automatic Spinal Deformity Detection Based on Neural Network

We propose a technique for automatic spinal deformity detection method from moire topographic images. Normally the moire stripes show a symmetric pattern, as a human body is almost symmetric. According to the progress of the deformity of a spine, asymmetry becomes larger. Numerical representation of the degree of asymmetry is therefore useful in evaluating the deformity. Displacement of local centroids is evaluated statistically between the left-hand side and the right-hand side regions of the moire images with respect to the extracted middle line. The degree of the displacement learned by a neural network employing the back propagation algorithm. An experiment was performed employing 1,200 real moire images (600 normal and 600 abnormal) and 89% of the images were classified correctly by the NN.

Spinal Deformity Detection from Moire Topographic Image Based on Evaluating Asymmetric Degree

In order to check the presence of spinal deformity in the early stage, orthopedists have traditionally performed on children a painless examination called a forward- bending test in school screening. In forward-bending test, mainly medical doctor checks to see if one shoulder is lower than the other. But this test is neither reproductive nor objective. Moreover, the inspection takes much time when applied to medical examination in schools. To overcome these difficulties, a moire method has been proposed which takes moire topographic images of human subject backs and checks symmetry/asymmetry of the moire patterns in a twodimensional way on visual screening. In this paper, we propose a new technique for automatic detection of spinal deformity from moire topographic images. In the first stage, once the original moire image is fed into computer, the middle line of the subject’s back is extracted on the moire image by employing the approximate symmetry analysis. Regions of interest are then automatically selected on the moire image from its upper part to the lower part. Numerical representation of the degree of asymmetry is therefore useful in evaluating the deformity. Displacement of local centroids and difference of gray value are calculated between the left-hand side and the right-hand side regions of the moire images with respect to the extracted middle line. Extracted 4 feature vectors (mean value and standard deviation from the each displacement) from the left-hand side and right-hand side rectangle areas are applied to train the Neural Network (NN), Support Vector Machines (SVMs). In the final stage, normal and abnormal cases are classified by NN and SVM. An experiment was performed employing 1,200 real moire images based on NN and SVMs, and classification rates of 90.3% and 85.3% was achieved respectively.