CN104462763B - Medical image interpretation method and device based on intelligent speech recognition - Google Patents
Medical image interpretation method and device based on intelligent speech recognition Download PDFInfo
- Publication number
- CN104462763B CN104462763B CN201410614512.7A CN201410614512A CN104462763B CN 104462763 B CN104462763 B CN 104462763B CN 201410614512 A CN201410614512 A CN 201410614512A CN 104462763 B CN104462763 B CN 104462763B
- Authority
- CN
- China
- Prior art keywords
- medical image
- deciphering
- result
- read signal
- solution read
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a medical image interpretation method based on intelligent speech recognition. The method comprises the steps that a medical image file to be interpreted is received; interpretation signals input in a speech mode are collected and recognized, and interpretation results corresponding to the interpretation signals are matched in a preset speech recognition bank, wherein the interpretation signals are speech signals input when the medical image file to be interpreted is interpreted; an image interpretation report is generated; the image interpretation report is displayed. The invention further discloses a medical image interpretation device based on intelligent speech recognition. According to the method and device, a medical image expert does not need to perform manual input to generate the image interpretation report; by adopting the speech recognition mode, the operation of medical image interpretation is simplified, and the efficiency and accuracy of medical image interpretation are improved.
Description
Technical field
The present invention relates to Medical Imaging Technology field, more particularly to a kind of medical image deciphering recognized based on intelligent sound
Method and apparatus.
Background technology
Medical image reading system is that radiologist is used to check, understands image and generate the system for understanding report,
There is very important effect in Medical Imaging field.Through development for many years, the system has completed digitized comprehensively, greatly
The development of Medical Imaging is promoted.But, existing medical image reading system is in use or presence is following not
Foot：
1st, when image is checked, it is impossible to synchronous to write report, but to be switched to and report the interface write or enter in program
OK；Also, when report is write, need mode manually that content is word for word input in computer.For one
A little relatively conventional inspection results (such as a part no abnormality seen), it is still desirable to carry out loaded down with trivial details input, be so accomplished by disappearing
The suitable energy and time of consumption；
2nd, complex to the common operation of image, needing repeatedly to click on associated button or menu to realize, such as adjust window,
Survey length etc..For fault image, it is impossible to accurately jump to specified location, needed for needing from current location to translate into one by one
Image.
3rd, the input mode of traditional keyboard+mouse is difficult to the image diagosis operation that perfect reply becomes increasingly complex, easily
There is the phenomenon that maloperation causes diagosis efficiency to decline.
The content of the invention
Present invention is primarily targeted at simplify the operation that medical image is read, efficiency that image understands and accurately is improved
Property.
For achieving the above object, the present invention provides a kind of medical image deciphering method recognized based on intelligent sound, described
The medical image deciphering method recognized based on intelligent sound is comprised the following steps：
Receive medical image file to be understood；
The solution read signal being input into by voice mode is gathered, the solution read signal is recognized, in preset speech recognition library
The corresponding deciphering result of matching solution read signal；It is described solution read signal be to described when the medical image file understood is understood
The voice signal being input into；
Generate and report is understood according to the deciphering result image of matching；
Show that the image for generating understands report.
Preferably, the medical image deciphering method recognized based on intelligent sound also includes step：
The phase in presetting range with the medical image file similarity is matched in preset Medical imaging
Like medical image data, and the knot that the corresponding image of the similar medical image data understands report is input into by voice mode
By as solution read signal.
Preferably, the identification solution read signal, the corresponding solution of matching solution read signal in preset speech recognition library
The step of reading result includes：
When the deciphering signal includes the corresponding position of preset first voice mark and medical image file, described
Deciphering masterplate corresponding with the position is matched in preset speech recognition library, as normal result is understood；
When the solution read signal does not include preset voice mark, match in the preset speech recognition library described
The corresponding word of solution read signal, as abnormal deciphering result；
When the deciphering signal includes the second preset voice mark, institute is matched in default Medical imaging
The corresponding word of solution read signal is stated, as the deciphering result of reference.
Preferably, it is described to show that the image for generating includes the step of understanding report：
Result is understood to the deciphering result and/or reference of the normal deciphering result in image deciphering report and/or exception
It is identified, the normal deciphering result and/or reference solution for understanding result and/or exception is shown with preset display mode
Read result.
Preferably, the medical image deciphering method recognized based on intelligent sound also includes step：
The mark instructions being marked to medical image file by phonetic entry are received, according to the mark instructions labelling
Corresponding medical image file, for jumping to corresponding medical image file after receiving look-up command when searching.
Additionally, for achieving the above object, the present invention is also provided a kind of medical image recognized based on intelligent sound and understands dress
Put, the medical image solution read apparatus recognized based on intelligent sound are included：
Receiver module, for receiving medical image file to be understood；
Identification and matching module, for the solution read signal that collection is input into by voice mode, recognizes the solution read signal, pre-
The corresponding deciphering result of matching solution read signal in the speech recognition library put；The solution read signal is to the medical science shadow to be understood
The voice signal being input into when being understood as file；
Report generation module, for generating image according to the deciphering result of matching report is understood；
Display module, for showing that the image for generating understands report.
Preferably, the medical image solution read apparatus recognized based on intelligent sound are also included：
Reference result matching module, for matching in default Medical imaging and the medical image file phase
Like similar medical image data of the degree in presetting range, and the similar medical image data correspondence is input into by voice mode
Image understand report conclusion, as solution read signal.
Preferably, the identification and matching module specifically for：
When the deciphering signal includes the corresponding position of preset first voice mark and medical image file, described
Deciphering masterplate corresponding with the position is matched in preset speech recognition library, as normal result is understood；
When the solution read signal does not include preset voice mark, match in the preset speech recognition library described
The corresponding word of solution read signal, as abnormal deciphering result；
When the deciphering signal includes the second preset voice mark, institute is matched in default Medical imaging
The corresponding word of solution read signal is stated, as the deciphering result of reference.
Preferably, the display module specifically for：
Result is understood to the deciphering result and/or reference of the normal deciphering result in image deciphering report and/or exception
It is identified, the normal deciphering result and/or reference solution for understanding result and/or exception is shown with preset display mode
Read result.
Preferably, the medical image solution read apparatus recognized based on intelligent sound are also included：
Mark module, for receiving the mark instructions being marked to medical image file by phonetic entry, according to
The corresponding medical image file of the mark instructions labelling, for jumping to corresponding medical science after receiving look-up command when searching
Image file.
The present invention is being received after the medical image file understood, during medical image expert understands, collection
The solution read signal being input into by voice mode, identification solution read signal simultaneously generates image solution according to the corresponding result of understanding of solution read signal
Read the newspaper announcement, show that the image for generating understands report.It is manually entered without the need for medical image expert generating when image understands report, leads to
The mode of speech recognition is crossed, the operation of medical image reading is simplified, and improves the efficiency and accurately of medical image deciphering
Property.
Description of the drawings
Fig. 1 is schematic flow sheet of the present invention based on the medical image deciphering method first embodiment of intelligent sound identification；
Fig. 2 is the refinement schematic flow sheet of step S20 in Fig. 1；
Fig. 3 is schematic flow sheet of the present invention based on the medical image deciphering method second embodiment of intelligent sound identification；
Fig. 4 is that the present invention is illustrated based on the functional module of the medical image solution read apparatus first embodiment of intelligent sound identification
Figure；
Fig. 5 is the high-level schematic functional block diagram of identification and matching module in Fig. 4；
Fig. 6 is that the present invention is illustrated based on the functional module of the medical image solution read apparatus second embodiment of intelligent sound identification
Figure.
The realization of the object of the invention, functional characteristics and advantage will be described further referring to the drawings in conjunction with the embodiments.
Specific embodiment
It should be appreciated that specific embodiment described herein is not intended to limit the present invention only to explain the present invention.
The present invention provides a kind of medical image deciphering method recognized based on intelligent sound.
With reference to Fig. 1, Fig. 1 is flow process of the present invention based on the medical image deciphering method first embodiment of intelligent sound identification
Schematic diagram.
In one embodiment, the medical image deciphering method for being recognized based on intelligent sound is included：
Step S10, receives medical image file to be understood；
Step S20, gathers the solution read signal being input into by voice mode, identification solution read signal, in preset speech recognition
The corresponding deciphering result of matching solution read signal in storehouse；
Step S30, generates image and understands report according to the deciphering result of matching；
Step S40, shows that the image for generating understands report.
Medical image data to be understood is received, the medical image data may include that each position of patient body is corresponding and treat
The medical image file of deciphering, and the personal information of patient, such as name, age, sex, medical history information.
When medical image expert understands medical image file, hardware that can be by mike etc. for phonetic entry sets
The solution read signal of standby input speech signal form, gathers the solution read signal, and to the solution read signal speech recognition is carried out, and according to knowledge
The solution read signal not gone out, the matching deciphering result corresponding with solution read signal in preset speech recognition library, understanding result can wrap
Include the normal deciphering result for understanding result and exception.Then, according to the deciphering result for matching, generate and understand knot for showing
The image of fruit understands report.When image deciphering report is generated, can be by the name of received patient, age, sex, disease
The information such as history are embedded in a modular fashion the front portion that image understands report, and can be automatic at the end of image deciphering report
Generate timestamp；Meanwhile, medical image expert can sign the name of oneself in image understands report.In the present embodiment,
Generate image to understand after report, such as find that the deciphering result that image is understood in report is wrong, then the content of mistake can be deleted
Remove.
Generating after image understands report, showing that the image understands report, that is, showing medical image expert to receiving
The deciphering result that medical image file is understood, when it is normal to understand result, then shows that corresponding position is normal, works as solution
When reading result is abnormal, then specific abnormal conditions are shown according to the deciphering of medical image expert.
The present embodiment is being received after the medical image file understood, and during medical image expert understands, is adopted
Collect the solution read signal being input into by voice mode, identification solution read signal simultaneously generates image according to the corresponding result of understanding of solution read signal
Report is understood, shows that the image for generating understands report.It is manually entered without the need for medical image expert generating when image understands report,
By way of speech recognition, the operation of medical image reading is simplified, and improve the efficiency and standard of medical image deciphering
True property.
With reference to Fig. 2, Fig. 2 is the refinement schematic flow sheet of step S20 in Fig. 1.
In the above-described embodiments, step S20 is specifically included：
Step S21, gathers the solution read signal being input into by voice mode, is filtered and noise reduction process to solving read signal；
When medical image expert understands medical image file, collect after the solution read signal of voice mode input,
Voice signal original to this first is processed, and is such as filtered process and noise reduction process, excludes the interference of voice signal simultaneously
Improve signal to noise ratio.
Step S22, identification solution read signal, matching word corresponding with solution read signal, makees in preset speech recognition library
To understand result；
Speech recognition, and matching and solution read signal pair in preset speech recognition library are carried out to the solution read signal after process
The word answered, as deciphering result.In the present embodiment, conventional medical terminology, radiation are have collected in preset speech recognition library
The masterplate of technics and common deciphering result, it is generally the case that understand result for normal corresponding to masterplate.
Specifically, in the present embodiment, the solution read signal for such as receiving includes preset the first voice mark and medical science
The corresponding position of image file, the first voice mark works as solution as judging to understand whether result is normal Warning Mark
When read signal includes the first voice mark, the deciphering result for showing medical image file is normal, the first voice mark
The content word that is ordinarily selected to different from related medical terminology and radiologic term.For example, when the solution read signal for receiving
Including " click+pulmonary ", then the deciphering result for showing the medical image file corresponding to pulmonary be it is normal, now, can be in voice
Matching template corresponding with pulmonary in identification storehouse, for example, match " pulmonary's no abnormality seen " and understand result as normal.
The solution read signal for such as receiving includes the second preset voice mark, and the second voice mark is understood as judgement
As a result it is whether the Warning Mark of the decipherings result of reference, i.e., when solution read signal includes the second voice mark, represents this
When need to match the deciphering result of reference corresponding with medical image file, the content of the second voice mark may be configured as " reference "
Etc. the word that can substantially determine that its deciphering result for indicating is the deciphering result for referring to.For example, when the solution read signal for receiving
For the conclusion that " reference " plus specific image understand report, then matching and image understand the conclusion pair of report in speech recognition library
The word answered, using the word as the deciphering result for referring to.
The solution read signal for such as receiving does not include preset voice mark, that is, solve and do not include in read signal the first voice mark
With the second voice mark, then can determine whether that it is not normally to understand result to understand result, now, recognize complete solution read signal, and
Further corresponding word is matched in speech recognition library according to solution read signal.Now, the solution read signal pair for such as matching
The word answered is unrelated with the deciphering result to medical image file, then show that the solution read signal for receiving is invalid signal；Phase
Instead, then show that it is abnormal deciphering result to solve the corresponding result of understanding of read signal.For example, the solution read signal for identifying is " two lungs
It can be seen that several soft-tissue density's stoves, form is consistent with metastatic carcinoma ", then corresponding word is matched in speech recognition library, by this
Word is used as abnormal deciphering result.
In the present embodiment, deciphering result corresponding with solution read signal is not such as matched in speech recognition library, then show this
Be not stored with corresponding word in speech recognition library, and now, medical image expert can be corresponding by keyboard or other modes input
Deciphering result understand in report to image, and the deciphering result of input is added into speech recognition library, in case connecing in next time
When receiving same or analogous solution read signal, corresponding deciphering result is matched from speech recognition library.
On the basis of medical image deciphering method first embodiment of the invention described above based on intelligent sound identification, second
In embodiment, the method also includes：
Similar doctor of the matching to medical image file similarity in presetting range in preset Medical imaging
Image data is learned, and the conclusion that the corresponding image of similar medical image data understands report is input into by voice mode, as solution
Read signal.
In the present embodiment, preset a Medical imaging, store in the Medical imaging it is conventional it is various just
Normal or abnormal medical image data can be being received after the medical image file understood for reference, can first to medical science
Medical image data in the medical image data and Medical imaging of image file carries out similarity mode, works as similarity
When in preset scope, the medical image data in Medical imaging is obtained as similar medical image data, it is preset
Scope can for example be set greater than 80% according to the type of medical image file and the self-defined setting of region of interest.
After getting similar medical image data, according to the corresponding medical image file of the similar medical image data and shadow
Corresponding relation as understanding the conclusion of report, the image for obtaining the medical image file understands the conclusion of report, and medical image is special
Family can be input into the conclusion for understanding report as solution read signal by way of phonetic entry.Image is understood in the conclusion reported and is removed
Deciphering result including medical image expert to the medical image file of patient, may also include the treatment meanss and health as adopted
The information such as multiple situation, are generating when the deciphering of the medical image file understood is reported, can report same with understanding by these information
Shi Jinhang shows, so that doctor and patient refer to.
When understanding wait the medical image file understood, matching and the medical image file in Medical imaging
Medical image data of the similarity of medical image data in presetting range, obtains Medical imaging and its corresponding shadow
As understand report conclusion, using the conclusion by phonetic entry mode as understand signal input, for doctor and patient reference, make
Patient although it is understood that the therapeutic modality with the similar situation of own situation, so as to provide convenience for patient, make doctor according to
Past case provides reference for the therapeutic scheme of this case.
Based on above-mentioned first, second embodiment of the present invention, step S40 is specially：
Result is understood to the deciphering result and/or reference of the normal deciphering result in image deciphering report and/or exception
It is identified, the normal deciphering result and/or reference solution for understanding result and/or exception is shown with preset display mode
Read result.
After deciphering result corresponding with the position of medical image file is matched in preset speech recognition library, according to solution
It is normal or abnormal to read result, and to the normal deciphering result for understanding result and exception different marks are carried out, while can be with
Reference result is designated other marks that can be distinguished with normal or abnormal deciphering result, such as by different deciphering result marks
Different fonts or different colors are designated as, when showing, to be shown with preset display mode, for example, to exception
The word content of deciphering result show or highlighted in overstriking form.In the present embodiment, such as image is understood in report
Deciphering result understand result for normal, then can be after matching and representing deciphering result for normally deciphering masterplate, to this
Understand the corresponding word content of masterplate and be designated blue font, and normally show；Understanding the deciphering result in report such as image is
Abnormal deciphering result, then can be red by the word content identification after the abnormal corresponding word of deciphering result is matched
Color font, and when showing the word content of the red font is shown or highlighted in overstriking form, to remind trouble
Person notes.
Deciphering when image deciphering report is generated, to the normal deciphering result in image deciphering report and/or exception
As a result different marks are carried out, and the normal deciphering result for understanding result and/or exception is shown with preset display mode, made
The display for obtaining image deciphering report is relatively sharp, facilitates patient to read.
With reference to Fig. 3, Fig. 3 is flow process of the present invention based on the medical image deciphering method 3rd embodiment of intelligent sound identification
Schematic diagram.
Based on medical image deciphering method first embodiment of the present invention based on intelligent sound identification, in 3rd embodiment
In, the method also includes：
Step S50, receives the mark instructions being marked to medical image file by phonetic entry, according to the labelling
The corresponding medical image file of cue mark, for jumping to corresponding medical image text after receiving look-up command when searching
Part.
In the present embodiment, during medical image expert understands medical image file, due to each position, institute is right
The medical image file answered may have multiple, therefore, during deciphering, such as a certain medical image file is had a question,
Or be intended to the medical image file as emphasis, it is desirable to the medical science shadow is found after whole medical image files have been checked
As file and re-start deciphering, then medical image expert can the input marking instruction by way of phonetic entry, be such as input into
" emphasis 1 " or " labelling the 42nd ", so, after the mark instructions are received, just enters rower to corresponding medical image file
Note.Before medical image expert searches during the medical image file of institute's labelling, equally can be input into by way of phonetic entry
Look-up command, such as input " jumping to emphasis 1 " or " jumping to the 42nd ", just can find the corresponding medical image text of labelling
Part, and show the medical image file.Step S50 can understand whenever carrying out for medical image file, this reality
Apply example preferably to perform before execution step S30 in the above-described first embodiment, that is, starting to understand medical image file and show
Whenever the image of generation is understood before report, and corresponding medical image file is marked.
When the mark instructions that are marked to medical image file are received, according to the corresponding medical science of mark instructions labelling
Image file, can facilitate accurately jump to after look-up command is received corresponding medical image file, so as to further
The operation of medical image reading is simplified, and improves the efficiency and accuracy of medical image deciphering.
The present invention also provides a kind of medical image solution read apparatus recognized based on intelligent sound.
With reference to Fig. 4, Fig. 4 is function of the present invention based on the medical image solution read apparatus first embodiment of intelligent sound identification
Module diagram.
In one embodiment, the medical image solution read apparatus for being recognized based on intelligent sound are included：
Receiver module 10, for receiving medical image file to be understood；
Identification and matching module 20, for the solution read signal that collection is input into by voice mode, identification solution read signal, preset
Speech recognition library in the matching solution corresponding deciphering result of read signal；Solution read signal is that the medical image file for treating deciphering is carried out
The voice signal being input into during deciphering；
Report generation module 30, for generating image according to the deciphering result of matching report is understood；
Display module 40, for showing that the image for generating understands report.
Receiver module 10 receives medical image data to be understood, and the medical image data may include each portion of patient body
The corresponding medical image file to be understood in position, and the personal information of patient, such as name, age, sex, medical history information.
When medical image expert understands medical image file, hardware that can be by mike etc. for phonetic entry sets
The solution read signal of standby input speech signal form, identification and matching module 20 gathers the solution read signal, to the solution read signal language is carried out
Sound is recognized, and according to the solution read signal for identifying, matching is corresponding with solution read signal in preset speech recognition library understands knot
Really, understanding result may include normally to understand the deciphering result of result and exception.Then, according to the deciphering result for matching, report
Accuse generation module 30 to generate for showing that the image for understanding result understands report.When image deciphering report is generated, can be by institute
The information such as name, age, sex, the medical history of the patient for receiving are embedded in a modular fashion image and understand before report
Portion, and the end that can understand report in image automatically generates timestamp；Meanwhile, medical image expert can read the newspaper in image solution
The name of oneself is signed in announcement.
After image deciphering report is generated, display module 40 shows that the image understands report, that is, show medical image expert
The deciphering result understood to the medical image file for receiving, when it is normal to understand result, then shows corresponding position
For normal, when understand result for it is abnormal when, then specific abnormal conditions are shown according to the deciphering of medical image expert.The present embodiment
In, after image deciphering report is generated, such as finding that the deciphering result that image is understood in reporting is wrong, then the content of mistake can be entered
Row is deleted.
The present embodiment is being received after the medical image file understood, and during medical image expert understands, is adopted
Collect the solution read signal being input into by voice mode, identification solution read signal simultaneously generates image according to the corresponding result of understanding of solution read signal
Report is understood, shows that the image for generating understands report.It is manually entered without the need for medical image expert generating when image understands report,
By way of speech recognition, the operation of medical image reading is simplified, and improve the efficiency and standard of medical image deciphering
True property.
With reference to Fig. 5, Fig. 5 is the high-level schematic functional block diagram of identification and matching module in Fig. 4.
In the above-described embodiments, identification and matching module 20 includes：
Acquisition process submodule 21, for the solution read signal that collection is input into by voice mode, filters to solving read signal
Ripple and noise reduction process；
Identification submodule 22, for identification solution read signal, matching is corresponding with solution read signal in preset speech recognition library
Word, as understand result.
When medical image expert understands medical image file, collect after the solution read signal of voice mode input,
The voice signal original to this first of acquisition process submodule 21 is processed, and is such as filtered process and noise reduction process, is excluded
The interference of voice signal simultaneously improves signal to noise ratio.
Solution read signal after identification submodule 22 pairs is processed carries out speech recognition, and matches in preset speech recognition library
Word corresponding with solution read signal, as deciphering result.In the present embodiment, conventional doctor is have collected in preset speech recognition library
The masterplate of technics, radiologic term and common deciphering result, it is generally the case that understanding for normal corresponding to masterplate
As a result.
Specifically, in the present embodiment, the solution read signal for such as receiving includes preset the first voice mark and medical science
The corresponding position of image file, the first voice mark works as solution as judging to understand whether result is normal Warning Mark
When read signal includes the first voice mark, the deciphering result for showing medical image file is normal, the first voice mark
The content word that is ordinarily selected to different from related medical terminology and radiologic term.For example, when the solution read signal for receiving
Including " click+pulmonary ", then the deciphering result for showing the medical image file corresponding to pulmonary be it is normal, now, can be in voice
Matching template corresponding with pulmonary in identification storehouse, for example, match " pulmonary's no abnormality seen " and understand result as normal.
The solution read signal for such as receiving includes the second preset voice mark, and the second voice mark is understood as judgement
As a result be whether the Warning Mark of the decipherings result of reference, i.e., when solution read signal includes the second voice mark, during table this
When need to match the deciphering result of reference corresponding with medical image file, the content of the second voice mark may be configured as " reference "
Etc. the word that can substantially determine that its deciphering result for indicating is the deciphering result for referring to.For example, when the solution read signal for receiving
For the conclusion that " reference " plus specific image understand report, then matching and image understand the conclusion pair of report in speech recognition library
The word answered, using the word as the deciphering result for referring to.
The solution read signal for such as receiving does not include preset voice mark, that is, solve and do not include in read signal the first voice mark
With the second voice mark, then can determine whether that it is not normally to understand result to understand result, now, recognize complete solution read signal, and
Further corresponding word is matched in speech recognition library according to solution read signal.Now, the solution read signal pair for such as matching
The word answered is unrelated with the deciphering result to medical image file, then show that the solution read signal for receiving is invalid signal；Phase
Instead, then show that it is abnormal deciphering result to solve the corresponding result of understanding of read signal.For example, the solution read signal for identifying is " two lungs
It can be seen that several soft-tissue density's stoves, form is consistent with metastatic carcinoma ", then corresponding word is matched in speech recognition library, by this
Word is used as abnormal deciphering result.
In the present embodiment, deciphering result corresponding with solution read signal is not such as matched in speech recognition library, then show this
Be not stored with corresponding word in speech recognition library, and now, medical image expert can be corresponding by keyboard or other modes input
Deciphering result understand in report to image, and the deciphering result of input is added into speech recognition library, in case connecing in next time
When receiving same or analogous solution read signal, corresponding deciphering result is matched from speech recognition library.
On the basis of medical image solution read apparatus first embodiment of the invention described above based on intelligent sound identification, second
In embodiment, the device also includes：
Reference result matching module, for the matching in default Medical imaging and medical image file similarity
Similar medical image data in presetting range, and the corresponding image solution of similar medical image data is input into by voice mode
Read the newspaper the conclusion of announcement, as solution read signal.
In the present embodiment, preset a Medical imaging, store in the Medical imaging it is conventional it is various just
Normal or abnormal medical image data can be being received after the medical image file understood, reference result matching mould for reference
Block can carry out phase to the medical image data in the medical image data and Medical imaging of medical image file first
Like degree matching, when similarity is in preset scope, the medical image data in Medical imaging is obtained as similar
Medical image data, preset scope can for example set according to the type of medical image file and the self-defined setting of region of interest
It is set to more than 80%.
After getting similar medical image data, according to the corresponding medical image file of the similar medical image data and shadow
Corresponding relation as understanding the conclusion of report, the image for obtaining the medical image file understands the conclusion of report, and medical image is special
Family can be input into the conclusion for understanding report as solution read signal by way of phonetic entry.Image is understood in the conclusion reported and is removed
Deciphering result including medical image expert to the medical image file of patient, may also include the treatment meanss and health as adopted
The information such as multiple situation, are generating when the deciphering of the medical image file understood is reported, can report same with understanding by these information
Shi Jinhang shows, for patient's reference.
When understanding wait the medical image file understood, matching and the medical image file in Medical imaging
Medical image data of the similarity of medical image data in presetting range, obtains Medical imaging and its corresponding shadow
Conclusion as understanding report, using the conclusion by phonetic entry mode as signal input is understood, for patient's reference, make the patient can
The therapeutic modality with the similar situation of own situation is recognized, so as to provide convenience for patient.
Based on the embodiment of the invention described above first, second, display module 40 specifically for：
Result is understood to the deciphering result and/or reference of the normal deciphering result in image deciphering report and/or exception
It is identified, the normal deciphering result and/or reference solution for understanding result and/or exception is shown with preset display mode
Read result.
After deciphering result corresponding with the position of medical image file is matched in preset speech recognition library, according to solution
It is normal or abnormal to read result, and to the normal deciphering result for understanding result and exception different marks are carried out, while can be with
Reference result is designated other marks that can be distinguished with normal or abnormal deciphering result, such as by different deciphering result marks
Different fonts or different colors are designated as, when showing, to be shown with preset display mode, for example, to exception
The word content of deciphering result show or highlighted in overstriking form.In the present embodiment, such as image is understood in report
Deciphering result understand result for normal, then can be after matching and representing deciphering result for normally deciphering masterplate, to this
Understand the corresponding word content of masterplate and be designated blue font, and normally show；Understanding the deciphering result in report such as image is
Abnormal deciphering result, then can be red by the word content identification after the abnormal corresponding word of deciphering result is matched
Color font, and when showing the word content of the red font is shown or highlighted in overstriking form, to remind trouble
Person notes.
Deciphering when image deciphering report is generated, to the normal deciphering result in image deciphering report and/or exception
As a result different marks are carried out, and the normal deciphering result for understanding result and/or exception is shown with preset display mode, made
The display for obtaining image deciphering report is relatively sharp, facilitates patient to read.
With reference to Fig. 6, Fig. 6 is function of the present invention based on the medical image solution read apparatus 3rd embodiment of intelligent sound identification
Module diagram.
Based on medical image solution read apparatus first embodiment of the present invention based on intelligent sound identification, in 3rd embodiment
In, the device also includes：
Mark module 50, for receiving the mark instructions being marked to medical image file by phonetic entry, root
According to the corresponding medical image file of the mark instructions labelling, for jumping to corresponding doctor after receiving look-up command when searching
Learn image file.
In the present embodiment, during medical image expert understands medical image file, due to each position, institute is right
The medical image file answered may have multiple, therefore, during deciphering, such as a certain medical image file is had a question,
Or be intended to the medical image file as emphasis, it is desirable to the medical science shadow is found after whole medical image files have been checked
As file and re-start deciphering, then medical image expert can the input marking instruction by way of phonetic entry, be such as input into
" emphasis 1 " or " labelling the 42nd ", so, after the mark instructions are received, mark module 50 is just to corresponding medical image
File is marked.Before medical image expert searches during the medical image file of institute's labelling, can equally pass through phonetic entry
Mode be input into look-up command, such as input " jumping to emphasis 1 " or " jumping to the 42nd " just can find the corresponding doctor of labelling
Image file is learned, and shows the medical image file.
When the mark instructions that are marked to medical image file are received, according to the corresponding medical science of mark instructions labelling
Image file, can facilitate accurately jump to after look-up command is received corresponding medical image file, so as to further
The operation of medical image reading is simplified, and improves the efficiency and accuracy of medical image deciphering.
The preferred embodiments of the present invention are these are only, the scope of the claims of the present invention is not thereby limited, it is every using this
Equivalent structure or equivalent flow conversion that bright description and accompanying drawing content are made, or directly or indirectly it is used in other related skills
Art field, is included within the scope of the present invention.
Claims (6)
1. it is a kind of based on intelligent sound recognize medical image deciphering method, it is characterised in that it is described based on intelligent sound recognize
Medical image deciphering method comprise the following steps：
Receive medical image file to be understood；
The solution read signal being input into by voice mode is gathered, the solution read signal is recognized, is matched in preset speech recognition library
The corresponding deciphering result of solution read signal；The solution read signal is by described defeated when the medical image file understood is understood
The voice signal for entering；
The similar doctor in presetting range to the medical image file similarity is matched in preset Medical imaging
Image data is learned, and the conclusion that the corresponding image of the similar medical image data understands report is input into by voice mode, made
To solve read signal；
Described to recognize the solution read signal, the matching in preset speech recognition library solves the step of read signal is corresponding to understand result
Including following (1) to (3) step：
(1) when the deciphering signal includes the corresponding position of preset first voice mark and medical image file, described
Deciphering masterplate corresponding with the position is matched in preset speech recognition library, as normal result is understood；
(2) when the solution read signal does not include preset voice mark, match in the preset speech recognition library described
The corresponding word of solution read signal, as abnormal deciphering result；
(3) when the deciphering signal includes the second preset voice mark, institute is matched in default Medical imaging
The corresponding word of solution read signal is stated, as the deciphering result of reference；
Image is generated according to the deciphering result of matching and understands report；
Show that the image for generating understands report.
2. it is as claimed in claim 1 to be based on the medical image deciphering method that intelligent sound is recognized, it is characterised in that the display
The step of image for generating understands report includes：
The deciphering result of the normal deciphering result in report and/or exception is understood to image and/or is carried out with reference to result is understood
Mark, with preset display mode the normal deciphering result and/or reference deciphering knot for understanding result and/or exception is shown
Really.
3. it is as claimed in claim 1 to be based on the medical image deciphering method that intelligent sound is recognized, it is characterised in that described to be based on
The medical image deciphering method of intelligent sound identification also includes step：
The mark instructions being marked to medical image file by phonetic entry are received, it is corresponding according to the mark instructions labelling
Medical image file, for search when receive look-up command after jump to corresponding medical image file.
4. it is a kind of based on intelligent sound recognize medical image solution read apparatus, it is characterised in that it is described based on intelligent sound recognize
Medical image solution read apparatus include：
Receiver module, for receiving medical image file to be understood；
Identification and matching module, for the solution read signal that collection is input into by voice mode, recognizes the solution read signal, preset
The corresponding deciphering result of matching solution read signal in speech recognition library；The solution read signal is to the medical image text to be understood
The voice signal being input into when part is understood；
Reference result matching module, for matching in default Medical imaging and the medical image file similarity
Similar medical image data in presetting range, and the corresponding shadow of the similar medical image data is input into by voice mode
Conclusion as understanding report, as solution read signal；
The identification and matching module specifically for：When the deciphering signal includes the first preset voice mark and medical image text
During the corresponding position of part, deciphering masterplate corresponding with the position is matched in the preset speech recognition library, as normal
Result is understood, when the solution read signal does not include preset voice mark, in the preset speech recognition library institute is matched
The corresponding word of solution read signal is stated, as abnormal deciphering result, when the deciphering signal includes the second preset voice mark
When, the corresponding word of the solution read signal is matched in default Medical imaging, as the deciphering result of reference；
Report generation module, for generating image according to the deciphering result of matching report is understood；
Display module, for showing that the image for generating understands report.
5. it is as claimed in claim 4 to be based on the medical image solution read apparatus that intelligent sound is recognized, it is characterised in that the display
Module specifically for：
The deciphering result of the normal deciphering result in report and/or exception is understood to image and/or is carried out with reference to result is understood
Mark, with preset display mode the normal deciphering result and/or reference deciphering knot for understanding result and/or exception is shown
Really.
6. it is as claimed in claim 4 to be based on the medical image solution read apparatus that intelligent sound is recognized, it is characterised in that described to be based on
The medical image solution read apparatus of intelligent sound identification also include：
Mark module, for receiving the mark instructions being marked to medical image file by phonetic entry, according to the mark
The corresponding medical image file of note cue mark, for jumping to corresponding medical image after receiving look-up command when searching
File.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410614512.7A CN104462763B (en) | 2014-11-04 | 2014-11-04 | Medical image interpretation method and device based on intelligent speech recognition |
| PCT/CN2014/090864 WO2016070445A1 (en) | 2014-11-04 | 2014-11-12 | Medical image interpretation method and system based on intelligent speech recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410614512.7A CN104462763B (en) | 2014-11-04 | 2014-11-04 | Medical image interpretation method and device based on intelligent speech recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104462763A CN104462763A (en) | 2015-03-25 |
| CN104462763B true CN104462763B (en) | 2017-05-03 |
Family
ID=52908792
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410614512.7A Expired - Fee Related CN104462763B (en) | 2014-11-04 | 2014-11-04 | Medical image interpretation method and device based on intelligent speech recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104462763B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107506604A (en) * | 2017-09-11 | 2017-12-22 | 深圳市前海安测信息技术有限公司 | Image recognition system and method based on artificial intelligence |
| CN113450897B (en) * | 2021-07-08 | 2023-07-14 | 思必驰科技股份有限公司 | Medical image film reading system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1758257A (en) * | 2005-09-27 | 2006-04-12 | 叶声 | Medical information management and search system using medical image as centre |
| CN103839211A (en) * | 2014-03-23 | 2014-06-04 | 合肥新涛信息科技有限公司 | Medical history transferring system based on voice recognition |
-
2014
- 2014-11-04 CN CN201410614512.7A patent/CN104462763B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1758257A (en) * | 2005-09-27 | 2006-04-12 | 叶声 | Medical information management and search system using medical image as centre |
| CN103839211A (en) * | 2014-03-23 | 2014-06-04 | 合肥新涛信息科技有限公司 | Medical history transferring system based on voice recognition |
Non-Patent Citations (2)
| Title |
|---|
| Voice activated command and control with Speech Recognition over WiFi;Ayres T et al.;《Science of Computer Programming》;20061231;第56卷(第2期);第109-126页 * |
| 放射科语音识别系统中环境自适应技术的研究与实现;李伟;《中国优秀硕士学位论文全文数据库 信息科技辑》;20131015(第10期);第33-34、40页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104462763A (en) | 2015-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104268603B (en) | Intelligent marking method and system for text objective questions | |
| CN109741806B (en) | Auxiliary generation method and device for medical image diagnosis report | |
| CN103530491B (en) | Apparatus and method for generating audit report | |
| CN109003648B (en) | Oral outpatient speech electronic medical record generation method and computer readable storage medium | |
| CN104850542B (en) | Non-audible voice input correction | |
| US11468989B2 (en) | Machine-aided dialog system and medical condition inquiry apparatus and method | |
| CN109478419B (en) | Automatic identification of salient discovery codes in structured and narrative reports | |
| CN107025174A (en) | For the method for the user interface abnormality test of equipment, device and readable storage media | |
| JP2013039230A (en) | Medical diagnosis support device and medical diagnosis support method | |
| CN103189858A (en) | Suggesting relevant terms during text entry | |
| CN107578798A (en) | The processing method and system of electronic health record | |
| RU2010128943A (en) | METHOD FOR DATA EXTRACTION FROM MEDICAL IMAGES DATA SET | |
| WO2015136404A1 (en) | System and method for scheduling healthcare follow-up appointments based on written recommendations | |
| CN111090641A (en) | Data processing method and device, electronic equipment and storage medium | |
| JP2019032704A (en) | Table data structuring system and table data structuring method | |
| US20210110559A1 (en) | Determining method, information processing apparatus, and recording medium | |
| WO2013160382A1 (en) | A system for reviewing medical image datasets | |
| CN104462763B (en) | Medical image interpretation method and device based on intelligent speech recognition | |
| CN108509034A (en) | Electronic device, information processing method and related product | |
| JP5789791B2 (en) | Similar case retrieval device and interpretation knowledge extraction device | |
| JP5589466B2 (en) | Information processing apparatus, program, recording medium, and information processing system | |
| US20230335261A1 (en) | Combining natural language understanding and image segmentation to intelligently populate text reports | |
| CN111144506A (en) | Ultrasound image-based hydatid identification method, storage medium and ultrasound equipment | |
| EP3719806A1 (en) | A computer-implemented method, an apparatus and a computer program product for assessing performance of a subject in a cognitive function test | |
| CN112819925A (en) | Method and device for processing focus labeling, electronic equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Hongzhi Inventor after: Xiao Yingfen Inventor after: Zheng Huihua Inventor after: Tang Xiaolang Inventor after: Zhang Shidao Inventor after: Hou Yunchao Inventor after: Zhao Xueming Inventor after: Zhang Guanjing Inventor after: Ge Xinke Inventor after: Wang Hairong Inventor after: Zhang Shaopeng Inventor after: Fang Jingfang Inventor after: Zhou Liang Inventor after: Jiang Xingju Inventor after: Liu Yi Inventor before: Zhang Guanjing Inventor before: Xu Zhiyan Inventor before: Zhou Liang Inventor before: Xu Juhong Inventor before: Fushizhu Inventor before: Liang Haoyuan Inventor before: Wang Jiayong Inventor before: Mao Kanlang Inventor before: Jiang Xingju Inventor before: Yang Qinglan Inventor before: Liu Yi Inventor before: Chen Xingming Inventor before: Xiao Yingfen Inventor before: He Xiaoxia Inventor before: Wu Binxia Inventor before: Zheng Huihua Inventor before: Tang Xiaolang Inventor before: Zhang Shidao Inventor before: Li Xiaoyun Inventor before: Hou Yunchao Inventor before: Zhao Xueming Inventor before: Zhao Xuezhu Inventor before: Ge Xinke Inventor before: Liu Guoyong Inventor before: Fu Dachu Inventor before: Ji Xuejian Inventor before: Zeng Chenduo Inventor before: Wang Hairong Inventor before: Zhang Shaopeng Inventor before: Fang Jingfang Inventor before: Cheng Jinjing Inventor before: Liang Yanni Inventor before: Zhou Rong |
|
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: Room 1501-1504, block a, block 9, block C, Wanke Yuncheng phase III, liuxinsi Road, Xili community, Nanshan District, Shenzhen City, Guangdong Province Patentee after: Shenzhen Anyuan Health Information Technology Co.,Ltd. Patentee after: E-TECHNO INFORMATION TECHNOLOGIES Co.,Ltd. Address before: 518057 Guangdong city of Shenzhen province Nanshan District Hing Road two No. 6 Wuhan University B815 Shenzhen research building (real Qianhai settled in Shenzhen City, Secretary of Commerce Co. Ltd.) Patentee before: ANYCHECK INFORMATION TECHNOLOGIES Co.,Ltd. Patentee before: E-TECHNO INFORMATION TECHNOLOGIES Co.,Ltd. |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170503 Termination date: 20211104 |