Invented by Qi Song, Feng Gao, Hanbo Chen, Shanhui Sun, Junjie Bai, Zheng Te, Youbing YIN, Keya Medical Technology Co Ltd

The market for system and method of generating and editing diagnostic reports based on images of medical objects has witnessed significant growth in recent years. This technology has revolutionized the field of medical imaging by streamlining the process of generating accurate and detailed diagnostic reports. Medical imaging plays a crucial role in the diagnosis and treatment of various medical conditions. It involves capturing images of internal body structures using techniques such as X-rays, magnetic resonance imaging (MRI), computed tomography (CT), and ultrasound. These images provide valuable insights into the patient’s condition, helping physicians make informed decisions about their treatment. Traditionally, the process of generating diagnostic reports based on medical images was time-consuming and prone to errors. Radiologists had to manually analyze the images, interpret the findings, and then draft a report. This process often led to delays in diagnosis and treatment, as well as potential inaccuracies in the reports. However, with the advent of systems and methods for generating and editing diagnostic reports based on medical images, this process has become more efficient and accurate. These systems utilize advanced algorithms and artificial intelligence (AI) to analyze the images and automatically generate preliminary reports. Radiologists can then review and edit these reports, adding their expertise and insights. One of the key advantages of this technology is its ability to improve the speed and efficiency of diagnosis. By automating the initial analysis of medical images, radiologists can focus on reviewing and interpreting the findings, leading to faster and more accurate diagnoses. This is particularly crucial in emergency situations where timely diagnosis can be a matter of life and death. Moreover, these systems also help in reducing errors and inconsistencies in diagnostic reports. The AI algorithms used in these systems are trained on vast amounts of medical data, enabling them to detect even subtle abnormalities in the images. This enhances the accuracy of the reports and reduces the chances of misdiagnosis. The market for systems and methods of generating and editing diagnostic reports based on medical images is expected to witness significant growth in the coming years. The increasing prevalence of chronic diseases, advancements in medical imaging technology, and the growing demand for faster and more accurate diagnoses are driving the adoption of these systems. Furthermore, the integration of these systems with electronic health records (EHR) and picture archiving and communication systems (PACS) is further fueling market growth. This integration allows for seamless sharing of images and reports between healthcare providers, improving collaboration and patient care. However, there are certain challenges that need to be addressed for the widespread adoption of these systems. One of the key challenges is ensuring the privacy and security of patient data. As these systems rely on AI algorithms, there is a need to ensure that patient data is protected and used in compliance with privacy regulations. In conclusion, the market for systems and methods of generating and editing diagnostic reports based on medical images is witnessing significant growth. This technology has the potential to revolutionize the field of medical imaging by improving the speed, accuracy, and efficiency of diagnosis. As advancements continue to be made in AI and medical imaging, we can expect further innovations in this market, leading to improved patient outcomes and healthcare delivery.

The Keya Medical Technology Co Ltd invention works as follows

Embodiments” of the disclosure are systems and methods of generating a document based on an image medical of a patient. A typical system comprises a communication device configured to receive a medical image captured by an image capture device. At least one processor may be included in the system. The processor is programmed to determine automatically keywords from the natural language description generated by a learning system of the medical picture. The processor is also configured to create a report that describes the medical image of a patient using the keywords. The processor can also display the report.

Background for System and method of generating and editing diagnostic reports based on images of medical objects

Radiologists use medical images to diagnose diseases and observe abnormalities. Writing diagnosis report is also part of routine jobs for radiologists/clinicians. Medical diagnosis reports, for example, summarize and describe important findings from medical images like X-rays, Computed Tomography images (CT), Magnetic Resonance Imaging images (MRI) and ultrasound images. The medical diagnosis reports are considered an important part of a patient’s profile. The current processes for producing medical diagnosis reports are not efficient.

The current processes for producing medical diagnosis reports take a lot of time, primarily due to two factors. The current processes require a careful visual inspection of at least one medical image of a patient to be made before any findings are recorded. Medical images can be large in comparison to the small lesions. This means that searching for suspicious areas may take some time. Second, findings and preliminary conclusions might need to be manually formatted into reports that can only be written by medical professionals, such as radiologists and clinicians.

Embodiments” of the disclosure addresses the problems above by providing a system for diagnosis reports that can analyze medical images automatically, detect suspicious areas, and generate diagnose reports.

Embodiments” of the disclosure are a system that generates a report using a medical picture of a patient. The system comprises a communication device configured to receive a medical image captured by an image capture device. At least one processor is also included in the system. The system includes at least one processor. This processor is designed to determine automatically keywords from the natural language description generated by the application of a learning networks to the medical images. The processor is also configured to create a report that describes the medical image of a patient using the keywords. The processor can also display the report.

Embodiments” of the disclosure provide a method of generating a document based on an image of a medical patient. The method comprises receiving the medical images via a communication interface. An image acquisition device acquires the medical image. At least one processor determines keywords automatically from the natural language description generated by a learning system of the medical picture. The method includes also generating a report by the processor describing the patient’s medical image based on keywords. The method also includes displaying the report.

Embodiments” of the disclosure provide non-transitory, computer-readable media with instructions that, when executed, will cause one or multiple processors to execute a method of generating a document based on an image medical of a patient. The method involves receiving the medical images acquired by an imaging device. The method also includes automatically determining the keywords from the natural language description generated by applying an image learning network. The method includes also generating a report that describes the medical image based on keywords. The method also includes displaying the report.

It is understood that the general description above and the detailed description below are only exemplary and informative and do not restrict the invention as claimed.

Now, we will refer in detail to the exemplary examples that are shown in the drawings.” The same reference numbers are used to refer to similar or identical parts wherever possible.

Embodiments” of the present disclosure allow for automated analysis of medical pictures and generation of reports on medical diagnosis. A system configured according to embodiments of the disclosure can automatically analyze medical pictures, detect suspicious areas, and generate diagnostic reports. In certain embodiments, deep-learning image processing and natural-language processing backend processes may support the system. In certain embodiments, the system allows medical professionals to edit/correct algorithm-generated reports and add new findings using an interactive user interface. A radiologist/clinician may, for example, manually type or dictate edits/corrections through the interactive user interface. By selecting medical keywords, a doctor can filter out a sentence from the generated report. In certain embodiments, systems configured according to embodiments of the disclosure can generate descriptions on a finer level in addition providing diagnosis reports based on whole images. A user can select one or several regions of interest from one or more images of medical care, and the system will automatically generate a diagnosis report based on the selected region. The systems and methods described in this disclosure are intended to reduce the time that medical professionals spend with each patient, and improve the efficiency of diagnosis.

In some embodiments, the system configured according to embodiments of this disclosure can support the automatic or semi-automatic creation of medical reports both for the whole image (or multiple images taken of the same patient) and/or a specific region of interest. The reports can include descriptions of observations. Images related to observations may be included in the reports.

In some embodiments, the system configured according to embodiments of this disclosure can generate and display the keywords of the descriptions for the clinical observations. The system can provide an interface which allows the user to select content to be reported using keywords.

In some embodiments the descriptions and keywords can be generated interactively based on the images that the user has selected to view. If a user decides to tile images, the system can generate a description that describes the overall impression. The system can also generate a description if a user selects to view only a portion of a 3D image. The system can generate a description if the user zooms in and views an enlarged portion of an image.

In some embodiments the keywords and descriptions can be generated interactively by combining the annotation information that is available to the system. The system can include annotation information from a user when it generates the descriptions and keywords.

In some embodiments, descriptions and keywords can be generated interactively by combining the speech information that is available to the system. The system can include speech information, for example, when generating descriptions and keywords.

In some embodiments, the system configured according to embodiments of this disclosure can automatically detect if the recorded speech is complete (e.g. not just a list of keywords). If it is determined that the speech is a complete description the system can convert the recorded speech into text (e.g. using one or more speech-recognition techniques) and then add the converted text in the report.

In some embodiments, descriptions and keywords can be generated interactively by combining text data available to the system. The system can include text information such as keywords or sentences that a user has entered.

In some embodiments, the deep learning model background may support a system configured according to embodiments of this disclosure. The end-to -end deep learning background process can be configured to combine a convolutional neural net (CNN) for image processing, a recurrent network (RNN) for natural language processing and an attention processing.

In some embodiments, the system configured according to embodiments of the current disclosure may allow the user to add images related to the reports when they add descriptions.

In some embodiments, an interactive system configured in accordance with embodiments of the present disclosure may significantly reduce the amount of time and workload of radiologists/clinicians compared with those involved in the traditional image inspection/diagnosis report writing procedure.

FIG. According to embodiments, FIG. 1 shows a block diagram for an example diagnosis report generating device 100. According to the present disclosure, a diagnosis reporting system 100 can be configured in a way that generates a report on the basis of medical images 102 captured by an image capture device 101. According to the present disclosure diagnosis report generating device 100 can receive medical images from image acquisition device. Medical images 102 can also be stored in a database of images (not shown), and the diagnosis report generating device 100 could receive images 102 directly from that database. In certain embodiments, the medical images 102 can be either two-dimensional (2D), or three-dimensional images (3D). A 3D image may contain multiple 2D image slices. “In some embodiments, the medical images 102 can contain images in tile views or different cross-sectional view, such as sagittal views, coronal views, and transverse views.

In certain embodiments, the image acquisition device may acquire medical images 102 by using any suitable imaging modalities. These include, e.g. X-rays, optical tomography and fluorescence imaging.

Click here to view the patent on Google Patents.