System and method for generating and editing diagnosis reports based on medical images
Embodiments of the disclosure provide systems and methods for generating a report based on medical images of a patient. An exemplary system includes a communication interface configured to receive the medical images acquired by an image acquisition device. The system may further include at least one processor. The at least one processor is configured to receive a user selection of at least one medical image in at least one view. The at least one processor is further configured to automatically generate keywords describing the selected medical image based on a learning network including a convolutional neural network and a recursive neural network connected in series. The at least one processor is also configured to receive a keyword selection among the generated keywords and generate the report based on the keyword selection. The exemplary system additionally includes a display configured to display the selected medical image and the report.
1. A system for generating a report based on medical images of a patient, comprising:
a communication interface configured to receive the medical images acquired by an image acquisition device;
at least one processor configured to:
receive a user selection of at least one medical image in at least one view;
automatically generate keywords describing the selected medical image based on a learning network including a convolutional neural network and a recursive neural network connected in series;
receive a keyword selection among the generated keywords; and
generate the report based on the keyword selection; and
a display configured to display the selected medical image and the report.
2. The system of claim 1 , wherein the at least one processor is further configured to extract image features from the selected medical image using the convolutional neural network.
3. The system of claim 2 , wherein the at least one processor is further configured to generate a natural language description of the selected medical image based on the image features using the recursive neural network.
4. The system of claim 3 , wherein the keywords are selected from the natural language description of the selected medical image.
5. The system of claim 2 , wherein the at least one processor is further configured to:
determine attention weights indicating contributions of respective image features extracted from the selected medical image to each generated keyword; and
weight the extracted image features with the corresponding attention weights.
6. The system of claim 1 , wherein the at least one processor is further configured to train the learning network by optimizing a joint loss function of the convolutional neural network and the recursive neural network.
7. The system of claim 1 , wherein the user selection further includes a region of interest in the selected medical image, and the at least one processor is configured to automatically update the keywords to describe the selected region of interest based on the learning network.
8. The system of claim 1 , wherein the at least one processor is further configured to:
receive requirements associated with generating the keywords; and
input the requirements to the recursive neural network.
9. The system of claim 1 , further comprising a microphone configured to receive a voice input, wherein the at least one processor is further configured to adjust the generated keywords based on a content of the voice input.
10. The system of claim 1 , wherein the at least one processor is further configured to depict a drawing tool on the display, the drawing tool being configured to receive an annotation, wherein the at least one processor is further configured to adjust the generated keywords based on the annotation.
11. The system of claim 1 , wherein the at least one processor is further configured to:
receive a text input; and
adjust the generated keywords based on the text input.
12. A method for generating a report based on medical images of a patient, comprising:
receiving, by a communication interface, the medical images acquired by an image acquisition device;
receiving a user selection of at least one medical image in at least one view;
automatically generating, by at least one processor, keywords describing the selected medical image based on a learning network including a convolutional neural network and a recursive neural network connected in series;
receiving a keyword selection among the generated keywords;
generating, by the at least one processor, the report based on the keyword selection; and
displaying on a display the selected medical image and the report.
13. The method of claim 12 , wherein automatically generating the keywords further includes:
extracting image features from the selected medical image using the convolutional neural network;
generating a natural language description of the selected medical image based on the image features using the recursive neural network; and
selecting the keywords from the natural language description.
14. The method of claim 13 , further comprising:
determining attention weights indicating contributions of respective image features extracted from the selected medical image to each generated keyword; and
weighting the extracted image features with the corresponding attention weights.
15. The method of claim 12 , further comprising training the learning network by optimizing a joint loss function of the convolutional neural network and the recursive neural network.
16. The method of claim 12 , wherein the user selection further includes a region of interest in the selected medical image, and the method further comprises automatically updating the keywords to describe the selected region of interest based on the learning network.
17. The method of claim 12 , further comprising:
receiving requirements associated with generating the keywords; and
inputting the requirements to the recursive neural network.
18. The method of claim 12 , further comprising:
receiving a user input, including at least one of a voice input, an annotation, or a text input; and
adjusting the generated keywords based on the user input.
19. A non-transitory computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by at least one processor, performs a method for generating a report based on medical images of a patient, comprising:
receiving the medical images acquired by an image acquisition device;
receiving a user selection of at least one medical image in at least one view;
automatically generating keywords describing the selected medical image based on a learning network including a convolutional neural network and a recursive neural network connected in series;
receiving a keyword selection among the generated keywords;
generating the report based on the keyword selection; and
displaying the selected medical image and the report.
20. The non-transitory computer-readable medium of claim 19 , wherein automatically generating the keywords further includes:
extracting image features from the selected medical image using the convolutional neural network;
generating a natural language description of the selected medical image based on the image features using the recursive neural network; and
selecting the keywords from the natural language description.