Radiation image processing method utilizing neural networks

• US 5,553,159 A
• Filed: 04/10/1992
• Issued: 09/03/1996
• Est. Priority Date: 04/16/1991
• Status: Expired due to Fees

• Current Assignee: Fujifilm Corporation (Fujifilm Holdings Corporation)
• Original Assignee: Fuji Photo Film Co Limited (Fujifilm Holdings Corporation)
• Inventors: Takeo, Hideya
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Tran, Phuoc
• Application Number:
  US07/867,060
• Time in Patent Office: 1,607 Days
• Field of Search: 382/14, 382/15, 382/6, 382/38, 382/36, 382/128, 382/131, 382/132, 382/155, 382/156, 382/157, 382/158, 382/159, 364/413.13, 364/413.14, 364/413.17, 364/43.19, 364/413.23, 395/2.11, 395/2.41, 395/2.68, 395/2.69, 395/21, 395/22, 395/23, 395/24, 395/25, 395/27
• US Class Current: 382/156
• CPC Class Codes: G01T 1/2014 Reading out of stimulable sheets, e.g. latent image
• CPC Class Codes: G06T 2207/20084 Artificial neural networks [ANN]
• CPC Class Codes: G06T 5/00 Image enhancement or restoration
• CPC Class Codes: G06T 5/60 using machine learning, e.g. neural networks
• CPC Class Codes: G06V 30/194 References adjustable by an adaptive method, e.g. learning
• CPC Class Codes: Y10S 128/925 Neural network

Abstract

In a radiation image processing method utilizing a neural network, an image signal representing a radiation image is fed into a neural network, image processing is carried out on the image signal by the neural network, and an output representing the results of the image processing is obtained from the neural network. Image processing, with respect to the whole region of the radiation image, is carried out on the image signal by a first group of neurons of an intermediate layer of the neural network. Image processing, with respect to parts of the region of the radiation image, is carried out on the image signal by a second group of neurons of the intermediate layer of the neural network. Regardless of set values of initial conditions, the output of the neural network becomes converged to a global minimum corresponding to the stored information, and the results of operation obtained from the neural network are not trapped at a local minimum.

Claims

1. A radiation image processing method utilizing a neural network, in which an image signal, representing a radiation image, is fed into a neural network, image processing is carried out on the image signal by the neural network, an output, representing the results of the image processing, is obtained from the neural network;

wherein the improvement comprises the steps of;

(i) providing the neural network with;

     1) an input layer of neurons, said input layer of neurons being made up of clusters of neurons, each of said clusters of neurons corresponding to a different part of the radiation image,

     2) an intermediate layer of neurons, said intermediate layer of neurons being made up of a first group of neurons and a second group of neurons, and

     3) an output layer of neurons;

(ii) carrying out the image processing on said image signal, with respect to the whole region of said radiation image, using the first group of neurons of the intermediate layer of said neural network, each of said first group of neurons being connected to each of the neurons of the input layer, and

(iii) carrying out the image processing on said image signal, with respect to parts of the region of said radiation image, using the second group of neurons of the intermediate layer of said neural network, each of the second group of neurons being connected to each of the neurons of one of said clusters,

iv combining results of the image processing carried out in steps ii and iii above to obtain the output representing the results of the image processing.

2. A method as defined in claim 1 wherein said image processing is a filtering process.

3. A method as defined in claim 1 wherein said image processing is one of the determination of a subdivision pattern of radiation images, the shape and location of an irradiation field, an orientation in which an object was placed when the image of the object was recorded, a portion of an object of which the image was recorded, and an abnormal pattern appearing in a radiation image.

4. A method as defined in claim 1 wherein said radiation image has been stored on a stimulable phosphor sheet.

5. A method as defined in claim 4 wherein said image signal is obtained by exposing said stimulable phosphor sheet to stimulating rays, which cause said stimulable phosphor sheet to emit light in proportion to the amount of energy stored thereon during its exposure to radiation, and photoelectrically detecting the emitted light.

6. A method as defined in claim 5 wherein said stimulating rays are a laser beam.

7. A method as defined in claim 1 wherein said radiation image has been recorded on photographic film.