Nasopharynx cancer screening method based on sound signal, terminal device and medium
Through the nasopharyngeal carcinoma screening method based on sound signals, the one-dimensional convolutional neural network model is used to analyze the sound signals, which solves the problems of high trauma, high cost and low accuracy of the existing nasopharyngeal carcinoma screening methods, and achieves low-cost and efficient nasopharyngeal carcinoma screening.
Patent Information
- Application Number
- CN202411906780.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-23
- Publication Date
- 2025-04-18
AI Technical Summary
The existing nasopharyngeal carcinoma screening methods have the problems of high trauma, high cost, low accuracy and difficulty in achieving extensive screening, especially in primary medical institutions that are difficult to conduct early diagnosis.
The nasopharyngeal carcinoma screening method is adopted based on sound signals, and the screening of nasopharyngeal carcinoma is achieved by collecting multi-person voice signals, signal segmentation, standardizing processing, building a one-dimensional convolutional neural network model, and using terminal equipment for detection.
This method is low-cost, efficient and harmless to the human body. It can provide a feasible large-scale screening program for nasopharyngeal carcinoma, reducing the difficulty and cost of screening.
Smart Images
Figure CN119851699A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of disease screening, and in particular to a nasopharyngeal carcinoma screening method based on sound signals, a terminal device and a medium. Background Art
[0002] The gold standard for clinical diagnosis of NPC is tissue pathological biopsy under nasopharyngeal endoscopy. Due to the large trauma and difficulty in obtaining samples, and the reliance on the operator's experience and skills, about 10% of patients fail the first diagnosis, and patients need to face the decision of re-biopsy or multiple biopsies, which can easily delay the disease. In addition, nasopharyngeal biopsy is difficult to achieve widespread screening of NPC due to the expensive medical equipment and difficult operation, and it is even more difficult to carry out early diagnosis in primary medical units. The existing clinical screening of NPC is mainly based on the detection of EBV markers in blood samples, which relies on clinical nurses to collect blood samples, limiting its promotion and detection in the entire population, and its accuracy is low, and there are usually false negatives, which makes it difficult to achieve widespread screening of NPC. Therefore, it is necessary to develop a NPC screening technology that is easy to promote and apply in areas with a high incidence of NPC. Summary of the invention
[0003] In order to solve the above problems, the present invention proposes a nasopharyngeal carcinoma screening method based on sound signals, a terminal device and a medium.
[0004] The specific plan is as follows:
[0005] A method for screening nasopharyngeal carcinoma based on sound signals, comprising the following steps:
[0006] S1: Collect the sound signals of multiple people corresponding to the same content;
[0007] S2: Segment the sound signal, where each sound signal corresponds to a sound segment of a fixed length;
[0008] S3: standardize each sound clip;
[0009] S4: After labeling each intercepted segment based on whether it has nasopharyngeal cancer, a training set is constructed based on each sound segment after standardization;
[0010] S5: Build a nasopharyngeal carcinoma screening model based on a one-dimensional convolutional neural network and train the model using the training set;
[0011] S6: When it is necessary to detect whether the user has nasopharyngeal carcinoma, the user's voice signal is collected and processed in steps S2 and S3, and then input into the trained model, and the output of the model is used as the detection result.
[0012] Furthermore, before the sound signal is segmented, the sound signal is filtered to reduce environmental noise.
[0013] Furthermore, the filtering adopts a Butterworth filter with a passband of 20 Hz to 2000 Hz.
[0014] Furthermore, when the sound clip is standardized, the signal is converted to conform to a normal distribution with a mean of 0 and a standard deviation of 1.
[0015] Furthermore, the network structure of the model includes two convolutional pooling modules, three fully connected layers and a Softmax activation function; the two convolutional pooling modules each include three convolutional pooling units connected in sequence, and each convolutional pooling unit includes a one-dimensional convolutional layer and a pooling layer; the sound clips received by the model are respectively input into the two convolutional pooling modules, and the outputs of the two convolutional pooling modules are spliced and passed through three fully connected layers in sequence, and finally the judgment result of whether the patient has nasopharyngeal cancer is output through the Softmax activation function.
[0016] Furthermore, the number of neurons in the three fully connected layers are 64, 32, and 8, respectively.
[0017] Furthermore, in the three convolutional pooling units included in each convolutional pooling module, the one-dimensional convolutional layers in the three convolutional pooling units are respectively composed of 16 32×1 convolutional kernels, 32 16×1 convolutional kernels and 64 8×1 convolutional kernels.
[0018] A nasopharyngeal carcinoma screening terminal device based on sound signals includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned method of the embodiment of the present invention are implemented.
[0019] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described above in an embodiment of the present invention are implemented.
[0020] The present invention adopts the above technical solution and has the following beneficial effects:
[0021] (1) This invention proposes for the first time to realize nasopharyngeal carcinoma examination based on sound signals combined with deep learning methods;
[0022] (2) The nasopharyngeal carcinoma examination method proposed by the present invention is low-cost, highly efficient and harmless to the human body;
[0023] (3) Provide a feasible solution for large-scale screening of NPC. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 Shown is a flow chart of a method according to a first embodiment of the present invention.
[0025] Figure 2FIG. 4 is a schematic diagram of a voice signal collected in this embodiment.
[0026] Figure 3 FIG. 4 is a schematic diagram of the speech signal segmentation result in this embodiment.
[0027] Figure 4 Shown is a schematic diagram of the network structure of the model in this embodiment. DETAILED DESCRIPTION
[0028] To further illustrate various embodiments, the present invention provides drawings. These drawings are part of the disclosure of the present invention, which are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in this field should be able to understand other possible implementations and advantages of the present invention.
[0029] The present invention will now be further described with reference to the accompanying drawings and specific implementation methods.
[0030] Embodiment 1:
[0031] The embodiment of the present invention provides a method for screening nasopharyngeal carcinoma based on sound signals, such as Figure 1 As shown, the method comprises the following steps:
[0032] S1: Collect sound signals corresponding to the same content from multiple people.
[0033] In this embodiment, the same recording device is used to collect voice signals of multiple people. Considering the complexity of the voice content and the convenience of collection, the content of the voice signal is set to say the numbers 1-10 in sequence. The measured voice signal example is as follows: Figure 2 shown.
[0034] S2: Segment the sound signal, where each sound signal corresponds to a sound segment of a fixed length.
[0035] This embodiment further includes filtering the sound signal before signal segmentation, and using a bandpass filtering algorithm to reduce environmental noise. Specifically, a third-order Butterworth filter is selected, and the bandpass frequency band is 20 Hz to 2000 Hz.
[0036] The signal segmentation algorithm used in this embodiment first determines the length of the sound segment based on the data average slope characteristics and standard deviation value, and then determines the initial point of the sound segment according to the data volatility, and finally realizes the segmentation of the sound signal and the extraction of each sound segment. Figure 3 shown.
[0037] S3: performing standardization processing on each sound clip.
[0038] In this embodiment, when performing standardization processing on the sound clip, the signal is converted into a normal distribution with a mean of 0 and a standard deviation of 1.
[0039] S4: After labeling each captured segment based on whether the individual has nasopharyngeal cancer, a training set is constructed based on each standardized sound segment.
[0040] In this embodiment, the label of the sound clip corresponds to the diagnosis result of whether the person corresponding to the sound clip has nasopharyngeal carcinoma, that is, the sound clip label of the person with nasopharyngeal carcinoma is set to 1, otherwise it is 0, and the data set required for subsequent neural network training is established. Furthermore, in order to prevent the neural network from overfitting, the order of samples in the data set needs to be disrupted. Finally, 90% of the data is set as the training set, and 10% of the data is set as the test set, of which 20% of the training set is set as the validation set for parameter adjustment.
[0041] S5: Construct a nasopharyngeal carcinoma screening model based on a one-dimensional convolutional neural network and train the model using a training set.
[0042] The model of this embodiment mainly consists of an input layer, a one-dimensional convolution layer, a pooling layer, a fully connected layer, an output layer, etc. The model takes the sound clips of the person being tested as input and outputs the result of whether the person has nasopharyngeal cancer.
[0043] The model specifically includes two independent convolutional pooling modules. Each convolutional pooling module includes three convolutional pooling units connected in sequence. Each convolutional pooling unit includes a one-dimensional convolutional layer and a pooling layer. After three convolution calculations, the features in the signal are extracted respectively. Among the three convolutional pooling units included in each convolutional pooling module, the one-dimensional convolutional layer in the three convolutional pooling units is composed of 16 32×1 convolution kernels, 32 16×1 convolution kernels and 64 8×1 convolution kernels respectively. After the two convolutional pooling modules complete the feature extraction, they are spliced and calculated through three fully connected layers (the number of neurons is 64, 32, and 8 respectively). Finally, the classification task is completed through the Softmax function. The overall network structure is as follows: Figure 4 In addition, the relu activation function is used to implement nonlinear mapping, the Adam optimizer is used, the learning rate is set to 0.0001, the Dropout layer parameter is set to 0.25, and the number of iterations is set to 500.
[0044] S6: When it is necessary to detect whether the user has nasopharyngeal carcinoma, the user's voice signal is collected and processed in steps S2 and S3, and then input into the trained model, and the output of the model is used as the detection result.
[0045] Embodiment 2:
[0046] The present invention also provides a nasopharyngeal carcinoma screening terminal device based on sound signals, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the above-mentioned method embodiment of embodiment 1 of the present invention when executing the computer program.
[0047] Further, as an executable solution, the nasopharyngeal carcinoma screening terminal device based on sound signals can be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The nasopharyngeal carcinoma screening terminal device based on sound signals may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that the composition structure of the above-mentioned nasopharyngeal carcinoma screening terminal device based on sound signals is only an example of a nasopharyngeal carcinoma screening terminal device based on sound signals, and does not constitute a limitation on the nasopharyngeal carcinoma screening terminal device based on sound signals. It may include more or fewer components than the above, or a combination of certain components, or different components. For example, the nasopharyngeal carcinoma screening terminal device based on sound signals may also include input and output devices, network access devices, buses, etc., and the embodiments of the present invention do not limit this.
[0048] Further, as an executable solution, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the nasopharyngeal carcinoma screening terminal device based on sound signals, and uses various interfaces and lines to connect various parts of the entire nasopharyngeal carcinoma screening terminal device based on sound signals.
[0049] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the nasopharyngeal carcinoma screening terminal device based on sound signals by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.
[0050] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method in the embodiment of the present invention are implemented.
[0051] If the module / unit integrated in the nasopharyngeal carcinoma screening terminal device based on sound signals is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory) and software distribution medium, etc.
[0052] Although the present invention has been specifically shown and described in conjunction with the preferred embodiments, it should be understood by those skilled in the art that various changes may be made to the present invention in form and details without departing from the spirit and scope of the present invention as defined by the appended claims, all of which are within the scope of protection of the present invention.
Claims
1. A method for screening nasopharyngeal carcinoma based on sound signals, characterized in that: The following steps are involved: S1: Collect the sound signals of multiple people corresponding to the same content; S2: Segment the sound signal, where each sound signal corresponds to a sound segment of a fixed length; S3: standardize each sound clip; S4: After labeling each intercepted segment based on whether it has nasopharyngeal cancer, a training set is constructed based on each sound segment after standardization; S5: Build a nasopharyngeal carcinoma screening model based on a one-dimensional convolutional neural network and train the model using the training set; S6: When it is necessary to detect whether the user has nasopharyngeal carcinoma, the user's voice signal is collected and processed in steps S2 and S3, and then input into the trained model, and the output of the model is used as the detection result.
2. The method for nasopharyngeal carcinoma screening based on sound signals according to claim 1, characterized in that: Before the sound signal is segmented, the sound signal is filtered to reduce the environmental noise.
3. The method for nasopharyngeal carcinoma screening based on sound signals according to claim 2, characterized in that: The filtering adopts a Butterworth filter with a passband of 20Hz to 2000Hz.
4. The method for nasopharyngeal carcinoma screening based on sound signals according to claim 1, characterized in that: When normalizing a sound clip, the signal is transformed to conform to a normal distribution with a mean of 0 and a standard deviation of 1.
5. The method for nasopharyngeal carcinoma screening based on sound signals according to claim 1, characterized in that: The network structure of the model includes two convolutional pooling modules, three fully connected layers and a Softmax activation function; the two convolutional pooling modules each include three convolutional pooling units connected in sequence, and each convolutional pooling unit includes a one-dimensional convolutional layer and a pooling layer; the sound clips received by the model are input into the two convolutional pooling modules respectively, and the outputs of the two convolutional pooling modules are spliced and passed through three fully connected layers in sequence, and finally the judgment result of whether the patient has nasopharyngeal cancer is output through the Softmax activation function.
6. The method for nasopharyngeal carcinoma screening based on sound signals according to claim 5, characterized in that: The number of neurons in the three fully connected layers are 64, 32, and 8 respectively.
7. The method for nasopharyngeal carcinoma screening based on sound signals according to claim 1, characterized in that: Each convolutional pooling module includes three convolutional pooling units, and the one-dimensional convolutional layers in the three convolutional pooling units are composed of 16 32×1 convolution kernels, 32 16×1 convolution kernels and 64 8×1 convolution kernels respectively.
8. A nasopharyngeal carcinoma screening terminal device based on sound signals, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 7 when executing the computer program.
9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.