阅读说明：本技术 一种人工智能听诊器及建立病理性声音库的方法 (artificial intelligent stethoscope and method for establishing pathological sound library ) 是由 邱亚娟 于 2019-10-30 设计创作，主要内容包括：本发明公开了一种人工智能听诊器及建立病理性声音库的方法,包括本体、听诊头、扬声器、手持部、模式选择键和显示屏,所述扬声区设有扬声器,所述扬声区上方的本体上设有手持部,所述手持部上方的本体上设有按键区,所述按键区设有模式选择键,所述按键区顶部设有显示屏,所述本体内部设有信息采集部和数据处理部；本发明结构和方法能够满足医务人员实时监测分析患者的心音、呼吸音等病理性声音数据,患者在医院外感到不适时通过该人工智能听诊器进行初步筛查,能及时发现病情进行早期诊断,能及时发现不能表达患者,如儿童及昏迷患者的病情变化,并提示给医生,智能化分析病理性声音,能够将诊断结果直观显示,适用范围广。(The invention discloses artificial intelligent stethoscopes and a method for establishing a pathological sound library, which comprise a body, an auscultation head, a loudspeaker, a handheld part, a mode selection key and a display screen, wherein the loudspeaker is arranged in a loudspeaker area, the handheld part is arranged on the body above the loudspeaker area, a key area is arranged on the body above the handheld part, the key area is provided with the mode selection key, the display screen is arranged at the top of the key area, and an information acquisition part and a data processing part are arranged in the body.)

The artificial intelligence stethoscope is characterized by comprising a body (1), an auscultation head (2), a loudspeaker (3), a handheld part (4), a mode selection key (5) and a display screen (6), wherein the auscultation head (2) is arranged on the body (1) through threads, a sound raising area (7) is arranged on the body (1) above the auscultation head (2), the sound raising area (7) is provided with the loudspeaker (3), a knob control switch (11) is arranged on the side of the loudspeaker (3) , the handheld part (4) is arranged on the body (1) above the sound raising area (7), a key area is arranged on the body (1) above the handheld part (4), the key area is provided with the mode selection key (5), the display screen (6) is arranged at the top of the key area, an information acquisition part (8) and a data processing part (10) are arranged inside the body (1), the information acquisition part (8) is connected with the data processing part (10), and the data processing part (10) is connected with the display screen (6).

2. The kinds of artificial intelligence stethoscope according to claim 1, wherein the sections of said stethoscope head (2) and said display screen (6) are circular.

3. The kinds of artificial intelligence stethoscope according to claim 1, wherein the stethoscope head (2) and the stethoscope body (1) are made of stainless steel.

4. The kinds of artificial intelligence stethoscope according to claim 1, wherein said stethoscope head (2) is a stethoscope head (2) having two forms of a membrane type and a drum type, said stethoscope head (2) is a membrane type for listening to high-pitched sounds, and said stethoscope head (2) is a drum type for listening to low-pitched sounds.

5. The kinds of artificial intelligence stethoscope according to claim 1, wherein said hand-held portion (4) is provided with anti-slip threads (9).

6. The kinds of artificial intelligence stethoscope according to claim 1, wherein said mode selection key (5) comprises a heart sound key, a breath sound key, a bowel sound key, a fetal heart sound key, a blood vessel noise key, a switch key, a catalog key, a top-bottom key and a result key, and said heart sound key, breath sound key, bowel sound key, fetal heart sound key, blood vessel noise key, switch key, catalog key, top-bottom key and result key are annularly arranged in the key area.

7, method for creating pathological sound library by artificial intelligence stethoscope, which comprises the following steps:

step , inputting personal information of a patient with a disease by using a display screen (6), then using an information acquisition part (8) to acquire pathological sounds of the patient with the disease confirmed by two doctors together, converting the acquired pathological sounds of the patient with the disease confirmed into a spectrogram of a digital signal by using the information acquisition part (8), and using the display screen (6) to visually display in real time, intercepting abnormal areas of the spectrogram by using the display screen (6), performing repeated auscultation to form a sound data set with a label, then intercepting characteristic segments by using the sound data set with the label of a data processing part (10), and performing frequency domain acoustic characteristic analysis;

step two: collecting the sound of a registered patient by using sound collection equipment, classifying the collected sound, and constructing a basic sound data set;

judging the categories of the basic sound data set in the sound data set with the labels in the step to obtain estimated pathological sound categories;

fourthly, constructing a convolutional neural network model, and then training the convolutional neural network model by utilizing the sound data set with the label after the frequency domain acoustic feature analysis in the to obtain a target recognition model;

and step five, integrating the sound data set with the label in the step , the target recognition model in the step two and the estimated pathological sound category in the step three to establish a pathological sound library.

8. The method for building a pathological sound database of artificial intelligence stethoscopes of claim 7, wherein the pathological sounds collected in step of the diagnosed patient include heart sounds, breath sounds, fetal heart sounds, bowel sounds and vascular murmurs, wherein the heart sounds and breath sounds include normal breath sounds, normal heart sounds, humorous sounds, wet ralos sounds, mitral stenosis, mitral insufficiency, ventricular septal defect and atrial septal defect.

Technical Field

The invention relates to the field of medical instruments, in particular to artificial intelligent stethoscopes and a method for establishing a pathological sound library.

Background

In clinical medical work, a stethoscope is a diagnosis and treatment tool commonly used by clinicians, and the application range of the stethoscope comprises heart sound auscultation, respiratory system auscultation, bowel sound auscultation, fetal heart auscultation, blood vessel murmur auscultation and the like. The common stethoscope is divided into a sound pickup part (a chest piece), a conduction part (a rubber tube) and a listening part (an ear piece);

the stethoscope is complex to use and store, inconvenient to carry, and often slips off when being hung on a neck, auscultation sounds are greatly interfered by the surrounding environment and cannot be adjusted in sound magnitude, in the process of diagnosis and treatment, children or unconscious patients cannot accurately express uncomfortable symptoms, so that delayed diagnosis is caused, the frequencies of partial heart sounds and breath sounds are overlapped, and the heart sounds and the breath sounds cannot be effectively separated even if a high-pass or low-pass filter is adopted.

Disclosure of Invention

In order to solve the problems, the invention provides artificial intelligent stethoscopes and a method for establishing a pathological sound library, which can meet the requirement that medical staff monitor and analyze pathological sound data of patients such as heart sounds, respiratory sounds and the like in real time, the patients can be preliminarily screened by the artificial intelligent stethoscopes when feeling out of hospitals inefficacy, the patients can be timely found out to have early diagnosis, the changes of the patients which can not express the conditions of the patients such as children and unconscious patients can be timely found out, the changes of the conditions of the patients can be prompted to the doctors, the pathological sounds can be intelligently analyzed, the diagnosis result can be visually displayed, and the application range is .

The invention provides artificial intelligent stethoscopes, which comprise a body, a stethoscopic head, a loudspeaker, a handheld part, a mode selection key and a display screen, wherein the stethoscopic head is arranged on the body through threads, a sound raising area is arranged on the body above the stethoscopic head and is provided with the loudspeaker, a knob control switch is arranged on the side of the loudspeaker , the handheld part is arranged on the body above the sound raising area, a key area is arranged on the body above the handheld part, the key area is provided with the mode selection key, the top of the key area is provided with the display screen, an information acquisition part and a data processing part are arranged in the body, the information acquisition part is connected with the data processing part, and the data processing part is connected with the display screen.

the improvement is that the section of the stethoscope head and the display screen are both circular.

the improvement is that the body and the stethoscope head are made of stainless steel.

the improvement is that the stethoscope head is a membrane type stethoscope head for listening to high-pitched sounds and a drum type stethoscope head for listening to low-pitched sounds.

the improvement is that the handle is provided with non-slip threads.

the improvement is that the mode selection key comprises heart sound key, breath sound key, bowel sound key, fetal heart sound key, blood vessel noise key, switch key, catalog key, up-down key and result key, and the heart sound key, breath sound key, bowel sound key, fetal heart sound key, blood vessel noise key, switch key, catalog key, up-down key and result key are annularly arranged in the key area.

method for creating pathological sound library by artificial intelligence stethoscope, comprising the following steps:

step , inputting individual information of the patient with disease by using a display screen, then using an information acquisition part to acquire pathological sounds of the patient with disease confirmed by two doctors together, converting the acquired pathological sounds of the patient with disease confirmed into a spectrogram of a digital signal by the information acquisition part, and using the display screen to visually display in real time, intercepting abnormal areas of the spectrogram by using the display screen, performing repeated auscultation to form a sound data set with a label, then intercepting characteristic fragments of the sound data set with the label by using a data processing part, and performing frequency domain acoustic characteristic analysis;

step two: collecting the sound of a registered patient by using sound collection equipment, classifying the collected sound, and constructing a basic sound data set;

judging the categories of the basic sound data set in the sound data set with the labels in the step to obtain estimated pathological sound categories;

fourthly, constructing a convolutional neural network model, and then training the convolutional neural network model by utilizing the sound data set with the label after the frequency domain acoustic feature analysis in the to obtain a target recognition model;

and step five, integrating the sound data set with the label in the step , the target recognition model in the step two and the estimated pathological sound category in the step three to establish a pathological sound library.

A further improvement resides in collecting pathological sounds of the diagnosed patient in step including heart sounds, breath sounds, fetal heart sounds, bowel sounds and vascular murmurs, wherein the heart sounds and breath sounds include normal breath sounds, normal heart sounds, rale, moist rale, mitral stenosis, mitral insufficiency, ventricular septal defect and atrial septal defect.

The artificial intelligent stethoscope has the advantages that the artificial intelligent stethoscope can meet the requirement that medical staff monitor and analyze pathological sound data of patients such as heart sounds and breath sounds in real time, the patients can be preliminarily screened by the artificial intelligent stethoscope when feeling improper outside a hospital, the illness state can be timely found for early diagnosis, the changes of the illness state of the patients such as children and unconscious patients can be timely found by the artificial intelligent stethoscope, the changes of the illness state of the patients which cannot be expressed can be prompted to doctors, the pathological sounds can be intelligently analyzed, the diagnosis result can be visually displayed, the application range is , the operation is simple, and the artificial intelligent stethoscope is not limited to be provided for professional doctors.

Drawings

FIG. 1 is a schematic front view of the structure of the present invention;

FIG. 2 is a sectional view showing the internal structure of the present invention;

FIG. 3 is a schematic view of the working principle of the structure of the present invention;

fig. 4 is a flow chart of a pathological sound library establishment method according to the present invention.

Wherein: 1. a body; 2. an auscultation head; 3. a speaker; 4. a hand-held portion; 5. a mode selection key; 6. a display screen; 7. a sound raising area; 8. an information acquisition unit; 9. anti-skid threads; 10. a data processing unit; 11. the knob controls the switch.

Detailed Description

For the purpose of enhancing understanding of the present invention, the present invention will be described in further detail in with reference to the following examples, which are provided for illustration only and are not to be construed as limiting the scope of the present invention.

According to fig. 1, 2, 3, 4 show, kinds of artificial intelligence stethoscopes are proposed to this embodiment, including body 1, auscultation head 2, speaker 3, handheld portion 4, mode selection key 5 and display screen 6, be equipped with auscultation head 2 through the screw thread on the body 1, be equipped with vocal area 7 on the body 1 of auscultation head 2 top, vocal area 7 is equipped with speaker 3, speaker 3 side is equipped with knob control switch 11, be equipped with handheld portion 4 on the body 1 of vocal area 7 top, be equipped with the key area on the body 1 of handheld portion 4 top, the key area is equipped with mode selection key 5, be equipped with display screen 6 at key area top, the inside information acquisition portion 8 and the data processing portion 10 of being equipped with of body 1, data processing portion 8 connects data processing portion 10, data processing portion 10 connects display screen 6.

The section surfaces of the auscultation head 2 and the display screen 6 are both circular.

The body 1 and the auscultation head 2 are made of stainless steel.

The stethoscope head 2 is a stethoscope head 2 with two forms of a membrane type and a drum type, the stethoscope head 2 is a membrane type and is used for listening to high-pitched sound, and the stethoscope head 2 is a drum type and is used for listening to low-pitched sound.

The hand-held part 4 is provided with an anti-slip thread 9.

The mode selection key 5 comprises a heart sound key, a breath sound key, an bowel sound key, a fetal heart sound key, a blood vessel noise key, a switch key, a directory key, an up-down key and a result key, and the heart sound key, the breath sound key, the bowel sound key, the fetal heart sound key, the blood vessel noise key, the switch key, the directory key, the up-down key and the result key are annularly arranged in a key area.

method for creating pathological sound library by artificial intelligence stethoscope, comprising the following steps:

step , inputting individual information of a disease patient by using a display screen 6, and then using an information acquisition part 8 by two medical doctors together to acquire pathological sounds of the disease patient, wherein the pathological sounds comprise heart sounds, breath sounds, fetal heart sounds, bowel sounds and blood vessel murmurs, the heart sounds and the breath sounds comprise normal breath sounds, normal heart sounds, dry rales, wet rales, mitral stenosis, mitral insufficiency, ventricular septal defects and atrial septal defects, the information acquisition part 8 converts the acquired pathological sounds of the disease patient to be diagnosed into a spectrogram of a digital signal and visually displays the spectrogram in real time by using the display screen 6, the display screen 6 intercepts abnormal regions of the spectrogram repeatedly to form a sound data set with a label, and then uses a sound data set with the label of a data processing part 10 to intercept characteristic segments and performs frequency domain acoustic characteristic analysis;

step two: collecting the sound of a registered patient by using sound collection equipment, classifying the collected sound, and constructing a basic sound data set;

judging the categories of the basic sound data set in the sound data set with the labels in the step to obtain estimated pathological sound categories;

fourthly, constructing a convolutional neural network model, and then training the convolutional neural network model by utilizing the sound data set with the label after the frequency domain acoustic feature analysis in the to obtain a target recognition model;

and step five, integrating the sound data set with the label in the step , the target recognition model in the step two and the estimated pathological sound category in the step three to establish a pathological sound library.

When used in this embodiment, the display screen 6 is used to input personal information of the patient, such as: name, sex, age, hospitalization number, then determining an auscultation mode according to the illness condition of a patient, selecting by a mode selection key 5, displaying a corresponding auscultation part on a display screen 6 to prompt an operator, simultaneously playing normal heart sounds or breath sounds and the like of the selected mode by a loudspeaker 3, placing an auscultation head 2 on the auscultation part of the patient, playing auscultation sounds by the loudspeaker 3, adjusting the volume of the loudspeaker 3 by the operator by using a knob control switch 11 according to the condition, simultaneously recording acquired information and displaying the recorded information on the display screen 6 in a spectrogram form in real time by an information acquisition part 8, analyzing the acquired information by a data processing part 10 through a pre-established pathological sound library, finally displaying the diagnosis result of auscultation on the display screen 6, and playing back the acquired sounds according to the requirement by the operator to perform repeated aus.

The artificial intelligent stethoscope can meet the requirement that medical staff monitors and analyzes pathological sound data of patients such as heart sounds and breath sounds in real time, the patients feel inappropriate outside hospitals and are preliminarily screened by the artificial intelligent stethoscope, the illness state can be found in time for early diagnosis, the artificial intelligent stethoscope can find the illness state change of the patients such as children and unconscious patients in time, the change of the illness state of the patients can be expressed, the pathological sound is intelligently analyzed, the diagnosis result can be visually displayed, the application range is , the operation is simple, and the artificial intelligent stethoscope is not only provided for professional doctors.

The foregoing illustrates and describes the principles, general features, and advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.