Deep learning medical systems and methods for image reconstruction and quality evaluation

• US 10,074,038 B2
• Filed: 11/23/2016
• Issued: 09/11/2018
• Est. Priority Date: 11/23/2016
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Hsieh, Jiang, Avinash, Gopal, Sirohey, Saad, Wang, Xin, Yin, Zhye, De Man, Bruno
• Primary Examiner(s): GORADIA, SHEFALI DINESH
• Application Number:
  US15/360,742
• Publication Number:
  US 20180144214A1
• Time in Patent Office: 657 Days
• CPC Class Codes: G06F 18/2193 based on specific statistical tests
• CPC Class Codes: G06N 3/04 Architecture, e.g. interconnection topology
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 3/084 Backpropagation, e.g. using gradient descent
• CPC Class Codes: G06T 2207/10081 Computed x-ray tomography [CT]
• CPC Class Codes: G06T 2207/10088 Magnetic resonance imaging [MRI]
• CPC Class Codes: G06T 2207/10104 Positron emission tomography [PET]
• CPC Class Codes: G06T 2207/10108 Single photon emission computed tomography [SPECT]
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/30168 Image quality inspection
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06V 10/454 Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
• CPC Class Codes: G06V 10/993 Evaluation of the quality of the acquired pattern

Abstract

Methods and apparatus to automatically generate an image quality metric for an image are provided. An example method includes automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image. The example method includes computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image.

Claims

1. A method to automatically generate an image quality metric for an image, the method comprising:

automatically processing a first medical image using a deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image;

computing the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image; and

outputting the first medical image and the associated image quality metric.

2. The method of claim 1, wherein the image quality metric includes an image quality index.

3. The method of claim 1, wherein the image quality metric indicates a confidence level for making a decision based on the first medical image, wherein a level 3 indicates sufficient confidence to make the decision, a level 5 indicates a highest level of confidence to make the decision, and a level 1 indicates no confidence to make the decision.

4. The method of claim 1, wherein the image quality metric represents at least one of a spatial resolution, a noise, or a task-based image quality associated with the first medical image.

5. The method of claim 1, wherein the deployed learning network model includes at least one of a feature-based machine learning network, a convolutional neural network, or an auto-encoder neural network.

6. The method of claim 1, wherein the first medical image includes at least one of a two-dimensional image, a three-dimensional image, or a four-dimensional image.

7. The method of claim 1, wherein the first medical image includes at least one of a magnetic resonance image, a positron emission tomography image, a single photon emission computed tomography image, an x-ray image, a computed tomography image, an ultrasound image, or a tomosynthesis image.

8. The method of claim 1, wherein automatically processing includes processing at least a portion of the first medical image.

9. The method of claim 1, wherein the digital learning and improvement factory is to be updated based on the first medical image and the associated image quality metric.

10. The method of claim 1, wherein outputting further includes providing the first medical image as at least one of a human-viewable image or a machine-processable image.

11. An apparatus to automatically generate an image quality metric for a medical image, the apparatus comprising:

a processor and memory configured to implement a deployed learning network model, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image, the processor configured to at least;

automatically process a first medical image using the deployed learning network model to generate an image quality metric for the first medical image;

compute the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image; and

output the first medical image and the associated image quality metric.

12. The apparatus of claim 11, wherein the image quality metric includes an image quality index.

13. The apparatus of claim 11, wherein the image quality metric indicates a confidence level for making a decision based on the first medical image, wherein a level 3 indicates sufficient confidence to make the decision, a level 5 indicates a highest level of confidence to make the decision, and a level 1 indicates no confidence to make the decision.

14. The apparatus of claim 11, wherein the image quality metric represents at least one of a spatial resolution, a noise, or a task-based image quality associated with the first medical image.

15. The apparatus of claim 11, wherein the deployed learning network model includes at least one of a feature-based machine learning network, a convolutional neural network, or an auto-encoder neural network.

16. The apparatus of claim 11, wherein automatically processing includes processing at least a portion of the first medical image.

17. The apparatus of claim 11, wherein the digital learning and improvement factory is to be updated based on the first medical image and the associated image quality metric.

18. The apparatus of claim 11, wherein providing further includes providing the first medical image as at least one of a human-viewable image or a machine-processable image.

19. The apparatus of claim 11, wherein the image quality metric is generated for a local patch of the medical image.

20. A non-transitory computer readable medium comprising instructions which, when executed, cause a machine to at least implement a deployed learning network model, the machine configured to at least:

automatically process a first medical image using the deployed learning network model to generate an image quality metric for the first medical image, the deployed learning network model generated from a digital learning and improvement factory including a training network, wherein the training network is tuned using a set of labeled reference medical images of a plurality of image types, and wherein a label associated with each of the labeled reference medical images indicates a central tendency metric associated with image quality of the image;

compute the image quality metric associated with the first medical image using the deployed learning network model by leveraging labels and associated central tendency metrics to determine the associated image quality metric for the first medical image; and

output the first medical image and the associated image quality metric.