Diagnostic imaging apparatus and method

Abstract

A diagnostic imaging apparatus and method are provided. The diagnostic imaging apparatus includes a detection unit configured to detect a lesion from a medical image, an interpretation unit configured to acquire a shape feature value by interpreting a shape of the detected lesion in a frequency domain, and a diagnosis unit configured to determine whether the detected lesion is benign or malignant based on the acquired shape feature value.

1 . A diagnostic imaging apparatus, comprising:
 a detection unit configured to detect a lesion from a medical image;   an interpretation unit configured to
 acquire a shape feature value by interpreting a shape of the detected lesion in a frequency domain; and 
   a diagnosis unit configured to determine whether the detected lesion is benign or malignant by comparing the acquired shape feature value with a model.   
     
     
         2 . The diagnostic imaging apparatus of  claim 1 , wherein, to acquire the shape feature value, the interpretation unit is further configured to extract one or more feature points from boundaries of the detected lesion, generate a shape histogram based on the extracted feature points, and convert the shape histogram to the frequency domain. 
     
     
         3 . The diagnostic imaging apparatus of  claim 2 , wherein the interpretation unit is further configured to generate a multi-dimensional shape histogram based on the extracted feature points. 
     
     
         4 . The diagnostic imaging apparatus of  claim 3 , wherein the interpretation unit is further configured to generate the multi-dimensional shape histogram based on a polar diagram or a log-polar diagram. 
     
     
         5 . The diagnostic imaging apparatus of  claim 2 , wherein the interpretation unit is further configured to generate a Fourier coefficient by applying a Fourier transform to the shape histogram, and determine the Fourier coefficient as the shape feature value. 
     
     
         6 . The diagnostic imaging apparatus of  claim 1 , wherein the model is established by supervised learning. 
     
     
         7 . The diagnostic imaging apparatus of  claim 1 , wherein the medical image comprises a radiograph, a sonogram, a magnetic resonance imaging (MRI) image, a computed tomography (CT) image, or any combination thereof. 
     
     
         8 . The diagnostic imaging apparatus of  claim 2 , wherein the interpretation unit comprises a feature point extractor, a shape histogram generator, a Fourier transformer, and a shape feature value acquirer, the feature point extractor being configured to extract the feature points corresponding to boundaries of the detected lesion, the shape histogram generator being configured to generate the shape histogram based on the extracted feature points, the Fourier transformer being configured to transform the generated shape histogram into normalized Fourier coefficients to convert the generated shape histogram to the frequency domain, the shape feature value acquirer being configured to acquire the shape feature value from the normalized Fourier coefficients. 
     
     
         9 . The diagnostic imaging apparatus of  claim 8 , wherein the feature point extractor is further configured to sample pixels corresponding to the boundaries of the detected lesion to extract the feature points. 
     
     
         10 . The diagnostic imaging apparatus of  claim 8 , wherein the feature point extractor is further configured to extract feature points that are equally spaced from each other. 
     
     
         11 . The diagnostic imaging apparatus of  claim 8 , wherein the feature point extractor is further configured to extract a greater amount of feature points from a point on the boundaries of the lesion comprising a greater curvature than from a point on the boundaries of the lesion comprising a lesser curvature. 
     
     
         12 . The diagnostic imaging apparatus of  claim 8 , wherein the shape feature value acquirer is further configured to determine the normalized Fourier coefficients as the shape feature value to determine a shape of the detected lesion. 
     
     
         13 . A diagnostic imaging method, comprising:
 detecting, using a detection unit, a lesion from a medical image;   acquiring, using an interpretation unit, a shape feature value, the acquiring comprising interpreting a shape of the lesion in a frequency domain; and   determining, using a diagnosis unit, whether the detected lesion is benign or malignant by comparing the acquired shape feature value with a model.   
     
     
         14 . The diagnostic imaging method of  claim 13 , wherein the acquiring further comprises extracting one or more feature points from boundaries of the detected lesion, generating a shape histogram based on the extracted feature points, and converting the generated shape histogram to the frequency domain. 
     
     
         15 . The diagnostic imaging method of  claim 14 , wherein the converting of the generated shape histogram comprises transforming the generated shape histogram into normalized Fourier coefficients. 
     
     
         16 . The diagnostic imaging method of  claim 14 , wherein the acquiring further comprises determining the normalized Fourier coefficients as the shape feature value to determine a shape of the detected lesion. 
     
     
         17 . The diagnostic imaging method of  claim 14 , wherein the extracted feature points are equally spaced from each other. 
     
     
         18 . The diagnostic imaging method of  claim 14 , wherein a greater amount of the extracted feature points is from a point on the boundaries of the lesion comprising a greater curvature than from a point on the boundaries of the lesion comprising a lesser curvature. 
     
     
         19 . The diagnostic imaging apparatus of  claim 2 , wherein the interpretation unit is further configured to adjust a number of or an interval between the feature points to be extracted according to properties of the boundaries of the detected lesion.