Method, apparatus and related device for feature determination of laryngeal image

Abstract

The application provides a laryngeal image feature determination method and device and related equipment, which determines the area proportion feature of a target abnormal area based on the area of the target abnormal area and the size information of a narrow-band endoscopic image, obtains the first quantity proportion feature of abnormal microvessels in a microvessel segmentation map, the second quantity proportion feature of hollow microvessels in the abnormal microvessels, obtains the third quantity proportion feature of irregular shape microvessels in the microvessel segmentation map, obtains the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantization feature in the target abnormal area, and adds the area proportion feature, the first quantity proportion feature, the second quantity proportion feature, the third quantity proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantization feature to a preset laryngeal image feature set. The abnormal degree of the larynx of a patient is efficiently and accurately determined.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of auxiliary medical technology, in particular to a laryngeal image feature determination method and device and related equipment. BACKGROUND

[0002] In recent years, with the development of minimally invasive diagnosis and treatment technology, it has become a new trend in laryngology to quickly and accurately diagnose and cure various laryngeal diseases.

[0003] However, at present, laryngoscopy in clinical practice must rely on the experience and subjective judgment of clinicians to identify pictures in order to diagnose and treat laryngeal diseases. Because the levels of clinical doctors are uneven, the diagnosis and treatment plans for the same disease are quite different in different level hospitals or by different doctors, and the quality of laryngoscopy diagnosis and treatment is not optimistic. This situation seriously restricts the development of precise laryngoscopy diagnosis and treatment.

[0004] Therefore, how to efficiently and accurately determine the detection image of laryngeal diseases is a technical problem that needs to be solved in the current field of auxiliary medical technology. SUMMARY

[0005] The present application provides a laryngeal image feature determination method, device and related equipment, aiming to solve the technical problem of how to efficiently and accurately determine the detection image of laryngeal diseases.

[0006] In one aspect, the present application provides a laryngeal image feature determination method, which comprises:

[0007] segmenting a target abnormal area in a pre-acquired narrow band endoscopic image to obtain an abnormal area segmentation map and an area of the target abnormal area, the narrow band endoscopic image being a narrow band endoscopic image taken for the larynx of a patient;

[0008] determining an area proportion feature of the target abnormal area based on the area of the target abnormal area and size information of the narrow band endoscopic image;

[0009] segmenting a microvessel area in the abnormal area segmentation map to obtain a microvessel segmentation map;

[0010] obtaining a first number proportion feature of abnormal microvessels in the microvessel segmentation map, a second number proportion feature of hollow microvessels in the abnormal microvessels;

[0011] obtaining a third number proportion feature of irregular shape microvessels in the microvessel segmentation map;

[0012] obtaining an abnormal microvessel aggregation degree feature, a microvessel density feature and an abnormal microvessel color quantization feature in the target abnormal area;

[0013] The area ratio feature, the first quantity ratio feature, the second quantity ratio feature, the third quantity ratio feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantification feature are added to a preset laryngeal image feature set used to determine the abnormality degree of the larynx of the patient.

[0014] In a possible implementation of the present application, the first quantity ratio feature of the abnormal microvessel in the microvessel segmentation map is obtained by:

[0015] Each microvessel in the microvessel segmentation map is extracted to obtain a first microvessel set, and a first total number of microvessels in the first microvessel set is counted;

[0016] The abnormal microvessels in the first microvessel set are obtained to obtain a second microvessel set, and a second total number of abnormal microvessels in the second microvessel set is counted;

[0017] The second total number is compared with the first total number to obtain a first ratio, and the first ratio is taken as the first quantity ratio feature of the abnormal microvessel in the microvessel segmentation map.

[0018] In a possible implementation of the present application, the first microvessel set is obtained by extracting each microvessel in the microvessel segmentation map, including:

[0019] On the basis of the connected domain, a minimum circumscribed horizontal rectangular frame of each microvessel in the microvessel segmentation map is obtained;

[0020] And each connected domain in the microvessel segmentation map is segmented to obtain a plurality of sub-connected domains;

[0021] Based on the minimum circumscribed horizontal rectangular frame and the plurality of sub-connected domains, each microvessel in the microvessel segmentation map is obtained to obtain a first microvessel set.

[0022] In a possible implementation of the present application, the abnormal microvessel in the first microvessel set is obtained by:

[0023] A first diameter list set of each microvessel in the first microvessel set is obtained, the first diameter list set including a plurality of diameters of each microvessel, the plurality of diameters being arranged in a preset size order;

[0024] A first diameter mean value of the first diameter list set of each microvessel is calculated;

[0025] Based on the first diameter mean value, a second diameter list set greater than the first diameter mean value and a third diameter list set smaller than the first diameter mean value are filtered from the first diameter list set of each microvessel;

[0026] respectively calculating a second diameter mean value of the second diameter list set and a third diameter mean value of the third diameter list set;

[0027] comparing the second diameter mean value and the third diameter mean value to obtain a second ratio value;

[0028] based on the second ratio value and a preset ratio threshold value, determining each microvessel to obtain an abnormal microvessel in the first microvessel set.

[0029] In a possible implementation of the present application, the second number proportion feature of the hollow microvessel in the abnormal microvessel includes:

[0030] color-reversing the minimum circumscribed rectangle frame of the abnormal microvessel in the second microvessel set to obtain a first black-and-white reversed image;

[0031] based on a preset connected domain method, traversing the first black-and-white reversed image and counting the number of connected domains in the first black-and-white reversed image;

[0032] if the number of connected domains is greater than a preset connected domain threshold value, it is determined that the abnormal microvessel is a hollow microvessel, and a third total value of the hollow microvessel in the second microvessel set is counted;

[0033] comparing the third total value with the second total value to obtain a third ratio value, and taking the third ratio value as the second number proportion feature of the hollow microvessel in the abnormal microvessel.

[0034] In a possible implementation of the present application, the third number proportion feature of the irregular shape microvessel in the microvessel segmentation image includes:

[0035] color-reversing the minimum circumscribed rectangle frame of each microvessel in the first microvessel set to obtain a second black-and-white reversed image;

[0036] along the width and height of the minimum circumscribed rectangle frame of each microvessel point, a preset length is taken as a target interval, and a plurality of first straight lines and a plurality of second straight lines parallel to the width and height are made in the minimum circumscribed rectangle frame of each microvessel;

[0037] respectively counting the first intersection number and the second intersection number of the target region boundary line in the minimum circumscribed rectangle frame of each microvessel for the plurality of first straight lines and the plurality of second straight lines;

[0038] based on the first intersection number, the second intersection number, the target interval and the size information of the minimum circumscribed rectangle frame of each microvessel, calculating an irregular quantization value of each microvessel;

[0039] determine shape-irregular microvessels in the first microvessel set based on the irregular quantization value of each microvessel and a preset irregular quantization threshold, and count a fourth total number value of the shape-irregular microvessels in the first microvessel set;

[0040] compare the fourth total number value with the first total number value to obtain a fourth ratio value, and use the fourth ratio value as a third quantity proportion feature of the shape-irregular microvessels in the microvessel segmentation map.

[0041] In a possible implementation of the present application, the adding of the area proportion feature, the first quantity proportion feature, the second quantity proportion feature, the third quantity proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantization feature to the preset laryngeal image feature set comprises:

[0042] adding the area proportion feature, the first quantity proportion feature, the second quantity proportion feature, the third quantity proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantization feature to the preset laryngeal image feature set;

[0043] weighting and fitting all features in the laryngeal image feature set to obtain an abnormal degree coefficient of the larynx;

[0044] determining the abnormal degree of the larynx of the patient based on the abnormal degree coefficient and a preset abnormal degree threshold.

[0045] In another aspect, the present application provides a feature determination device for a laryngeal image, which comprises:

[0046] a first segmentation unit configured to segment a target abnormal region in a pre-acquired narrow-band endoscopic image to obtain an abnormal region segmentation map and an area of the target abnormal region, the narrow-band endoscopic image being a narrow-band endoscopic image taken for a larynx of a patient;

[0047] a first determination unit configured to determine an area proportion feature of the target abnormal region based on the area of the target abnormal region and size information of the narrow-band endoscopic image;

[0048] a second segmentation unit configured to segment a microvessel region in the abnormal region segmentation map to obtain a microvessel segmentation map;

[0049] a first acquisition unit configured to acquire a first quantity proportion feature of abnormal microvessels in the microvessel segmentation map and a second quantity proportion feature of hollow microvessels in the abnormal microvessels;

[0050] a second acquisition unit configured to acquire a third quantity proportion feature of shape-irregular microvessels in the microvessel segmentation map;

[0051] a third obtaining unit, configured to obtain an abnormal microvessel aggregation degree feature, a microvessel density feature, and an abnormal microvessel color quantification feature in the target abnormal area;

[0052] a first adding unit, configured to add the area ratio feature, the first quantity ratio feature, the second quantity ratio feature, the third quantity ratio feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantification feature to a preset laryngeal image feature set, the laryngeal image feature set being used to determine an abnormality degree of the larynx of the patient.

[0053] In a possible implementation of the present application, the first obtaining unit specifically comprises:

[0054] a first extracting unit, configured to extract each microvessel in the microvessel segmentation map to obtain a first microvessel set, and count a first total number of microvessels in the first microvessel set;

[0055] a second obtaining unit, configured to obtain abnormal microvessels in the first microvessel set to obtain a second microvessel set, and count a second total number of abnormal microvessels in the second microvessel set;

[0056] a third obtaining unit, configured to compare the second total number with the first total number to obtain a first ratio, and take the first ratio as a first quantity ratio feature of abnormal microvessels in the microvessel segmentation map.

[0057] In a possible implementation of the present application, the first extracting unit is specifically configured to:

[0058] obtain a minimum circumscribed horizontal rectangle frame of each microvessel in the microvessel segmentation map on the basis of a connected domain;

[0059] and segment each connected domain from the microvessel segmentation map to obtain a plurality of sub-connected domains;

[0060] obtain each microvessel in the microvessel segmentation map based on the minimum circumscribed horizontal rectangle frame and the plurality of sub-connected domains to obtain a first microvessel set.

[0061] In a possible implementation of the present application, the second obtaining unit is specifically configured to:

[0062] obtain a first diameter list set of each microvessel in the first microvessel set, the first diameter list set comprising a plurality of diameters of each microvessel, the plurality of diameters being arranged in a preset size order;

[0063] calculate a first diameter mean value of the first diameter list set of each microvessel;

[0064] filtering, from the first diameter list set of each microvessel, a second diameter list set greater than the first diameter mean and a third diameter list set less than the first diameter mean based on the first diameter mean;

[0065] respectively calculating a second diameter mean of the second diameter list set and a third diameter mean of the third diameter list set;

[0066] comparing the second diameter mean and the third diameter mean to obtain a second ratio;

[0067] based on the second ratio and a preset ratio threshold, determining each microvessel to obtain an abnormal microvessel in the first microvessel set.

[0068] In a possible implementation of the present application, the first acquisition unit is further configured to:

[0069] performing color inversion processing on the minimum circumscribed rectangle frame of the abnormal microvessel in the second microvessel set to obtain a first black-and-white inversion image;

[0070] based on a preset connected domain method, traversing the first black-and-white inversion image and counting a number of connected domains in the first black-and-white inversion image;

[0071] if the number of connected domains is greater than a preset connected domain threshold, determining that the abnormal microvessel is a hollow microvessel and counting a third total value of hollow microvessels in the second microvessel set;

[0072] comparing the third total value with the second total value to obtain a third ratio, and taking the third ratio as a second quantity ratio feature of hollow microvessels in the abnormal microvessel.

[0073] In a possible implementation of the present application, the second acquisition unit is specifically configured to:

[0074] performing color inversion processing on the minimum circumscribed rectangle frame of each microvessel in the first microvessel set to obtain a second black-and-white inversion image;

[0075] presetting a length as a target interval along the width and height of the minimum circumscribed rectangle frame of each microvessel point, and making a plurality of first straight lines and a plurality of second straight lines parallel to the width and height in the minimum circumscribed rectangle frame of each microvessel;

[0076] respectively counting a first intersection number and a second intersection number of the plurality of first straight lines and the plurality of second straight lines with a target region boundary line in the minimum circumscribed rectangle frame of each microvessel;

[0077] calculate an irregular quantization value of each microvessel based on the first intersection number, the second intersection number, the target interval, and size information of the minimum circumscribed rectangular frame of each microvessel;

[0078] determine a shape-irregular microvessel in the first microvessel set based on the irregular quantization value of each microvessel and a preset irregular quantization threshold, and count a fourth total value of the shape-irregular microvessel in the first microvessel set;

[0079] compare the fourth total value with the first total value to obtain a fourth ratio value, and use the fourth ratio value as a third quantity proportion feature of the shape-irregular microvessel in the microvessel segmentation map.

[0080] In a possible implementation of the present application, the first adding unit is specifically configured to:

[0081] add the area proportion feature, the first quantity proportion feature, the second quantity proportion feature, the third quantity proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantization feature to a preset laryngeal image feature set;

[0082] perform weighted fitting on all features in the laryngeal image feature set to obtain an abnormal degree coefficient of the larynx;

[0083] determine the abnormal degree of the larynx of the patient based on the abnormal degree coefficient and a preset abnormal degree threshold.

[0084] In another aspect, the present application also provides a computer device, which comprises:

[0085] one or more processors;

[0086] a memory; and

[0087] one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the feature determination method of the laryngeal image.

[0088] In another aspect, the present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is loaded by a processor to execute the steps in the feature determination method of the laryngeal image.

[0089] The feature determination method of the laryngeal image provided in the application, by segmenting the target abnormal area in the narrow band endoscope image obtained in advance, the abnormal area segmentation graph and the area of the target abnormal area are obtained, and the narrow band endoscope image is the narrow band endoscope image for shooting the larynx of the patient; based on the area of the target abnormal area and the size information of the narrow band endoscope image, the area proportion feature of the target abnormal area is determined; the microvessel area in the abnormal area segmentation graph is segmented to obtain the microvessel segmentation graph; the first number proportion feature of the abnormal microvessel in the microvessel segmentation graph, the second number proportion feature of the hollow microvessel in the abnormal microvessel; the third number proportion feature of the irregular shape microvessel in the microvessel segmentation graph is obtained; the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantization feature in the target abnormal area are obtained; the area proportion feature, the first number proportion feature, the second number proportion feature, the third number proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantization feature are added to the preset laryngeal image feature set, and the laryngeal image feature set is used to determine the abnormal degree of the larynx of the patient. Compared with the traditional method, in the case that the abnormal degree of the larynx of the patient cannot be determined efficiently and accurately, the application creatively proposes that the target abnormal area of the narrow band endoscope image of the larynx of the patient is segmented, then the target abnormal area is comprehensively analyzed to obtain the corresponding multiple feature indexes, and finally the multiple feature indexes are comprehensively analyzed, so that the abnormal degree of the larynx of the patient is determined efficiently and accurately. BRIEF DESCRIPTION OF DRAWINGS

[0090] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creating labor.

[0091] Figure 1 is a scene schematic diagram of the laryngeal image feature determination system provided by the embodiments of the application;

[0092] Figure 2 is an embodiment flowchart of the laryngeal image feature determination method provided in the embodiments of the application;

[0093] Figure 3 is a blood vessel quantization schematic diagram provided in the embodiments of the application;

[0094] Figure 4 is a blood vessel segmentation graph black and white reversal schematic diagram provided in the embodiments of the application;

[0095] Figure 5 is a blood vessel irregularity degree quantization schematic diagram provided in the embodiments of the application;

[0096] Figure 6 is an embodiment structure schematic diagram of the throat image feature determination device provided in the embodiments of the present application;

[0097] Figure 7 is an embodiment structure schematic diagram of the computer device provided in the embodiments of the present application. DETAILED DESCRIPTION

[0098] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0099] In the description of the present application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings, and are only for the purpose of facilitating the description of the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features limited by "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0100] In the present application, the word "exemplary" is used to mean "serving as an example, instance, or illustration." Any implementation described as "exemplary" in the present application is not necessarily to be construed as preferred or advantageous over other implementations. The following description is presented to enable any person skilled in the art to make and use the present application. In the following description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the present application. It should be apparent to one skilled in the art, however, that the present application can be practiced without using these specific details. In other instances, well-known structures and processes are not elaborated upon in order to avoid unnecessary detail, which can obscure the description of the present application. Accordingly, the present application is not intended to be limited by the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0101] Embodiments of the present application provide a laryngeal image feature determination method and device and related equipment, which are described in detail below.

[0102] As shown in Figure 1 , Figure 1 is a scene schematic diagram of a laryngeal image feature determination system provided by embodiments of the present application. The laryngeal image feature determination system can include a computer device 100, and the computer device 100 is integrated with a laryngeal image feature determination device, such as Figure 1 the computer device 100 in .

[0103] In embodiments of the present application, the computer device 100 is mainly used to segment a target abnormal area in a pre-acquired narrow-band endoscopic image to obtain an abnormal area segmentation map and an area of the target abnormal area. The narrow-band endoscopic image is a narrow-band endoscopic image taken for the larynx of a patient. Based on the area of the target abnormal area and the size information of the narrow-band endoscopic image, an area ratio feature of the target abnormal area is determined. The microvessel area in the abnormal area segmentation map is segmented to obtain a microvessel segmentation map. A first number ratio feature of abnormal microvessels in the microvessel segmentation map, a second number ratio feature of hollow microvessels in the abnormal microvessels, a third number ratio feature of irregular shape microvessels in the microvessel segmentation map, an abnormal microvessel aggregation degree feature, a microvessel density feature, and an abnormal microvessel color quantization feature in the target abnormal area are obtained. The area ratio feature, the first number ratio feature, the second number ratio feature, the third number ratio feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantization feature are added to a preset laryngeal image feature set, and the laryngeal image feature set is used to determine the abnormality degree of the larynx of the patient.

[0104] In embodiments of the present application, the computer device 100 can be a terminal or a server. When the computer device 100 is a server, it can be an independent server, or a server network or a server cluster composed of servers. For example, the computer device 100 described in embodiments of the present application includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets, or a cloud server constructed by a plurality of servers. The cloud server is constructed by a large number of computers or network servers based on cloud computing.

[0105] It can be understood that the computer device 100 in the embodiment of the application is a terminal, and the terminal used can be a device including receiving and transmitting hardware, that is, a device having receiving and transmitting hardware capable of performing bidirectional communication on a bidirectional communication link. Such a device can include a cellular or other communication device with a single-line display or a multi-line display or a cellular or other communication device without a multi-line display. The computer device 100 can be a desktop terminal or a mobile terminal, and the computer device 100 can also be one of a mobile phone, a tablet computer, a notebook computer, a medical auxiliary instrument, and the like.

[0106] Those skilled in the art can understand that Figure 1 The application environment shown in the above embodiment is only one application scenario of the scheme of the application, and does not limit the application scenario of the scheme of the application. Other application environments can include more or fewer computer devices than Figure 1 The above embodiment only shows one computer device, and it can be understood that the throat image feature determination system can also include one or more other computer devices, which are not limited here. Figure 1 The above embodiment only shows one computer device, and it can be understood that the throat image feature determination system can also include one or more other computer devices, which are not limited here.

[0107] In addition, as shown in the above embodiment Figure 1 The throat image feature determination system can also include a memory 200 for storing data, such as storing narrow-band endoscopic images of a patient's throat and throat image feature determination data, for example, throat image feature determination data when the throat image feature determination system is running.

[0108] It should be noted that Figure 1 The scene diagram of the throat image feature determination system shown in the above embodiment is only one example, and the throat image feature determination system and the scene described in the above embodiment are used to more clearly illustrate the technical scheme of the application, and do not limit the technical scheme provided by the application. Those skilled in the art can know that, with the evolution of the throat image feature determination system and the appearance of new business scenarios, the technical scheme provided by the application is also applicable to similar technical problems.

[0109] Next, the throat image feature determination method provided by the application is introduced.

[0110] In the embodiment of the throat image feature determination method of the application, the throat image feature determination device is taken as the execution subject. In order to simplify and facilitate the description, the execution subject will be omitted in the following method embodiments, and the throat image feature determination device is applied to a computer device.

[0111] Please refer to Figure 2 to Figure 7 , Figure 2For an embodiment flowchart of the throat image feature determination method provided in the embodiments of the present application, the throat image feature determination method comprises:

[0112] 201. Segment the target abnormal area in the pre-acquired narrow band endoscopic image to obtain an abnormal area segmentation map and an area of the target abnormal area.

[0113] The narrow band endoscopic image is a narrow band endoscopic image taken of the throat of a patient.

[0114] The narrow band endoscopic image of the throat of the patient can be acquired by a narrow band imaging endoscope (NBI), and the size information of the narrow band endoscopic image is specifically the width W and the height H of the narrow band endoscopic image. The NBI is a filter that filters out the wide band spectrum of the red, blue and green light waves emitted by the endoscope light source, leaving only the narrow band spectrum for the diagnosis of various diseases of the digestive tract.

[0115] In the embodiments of the present application, the target abnormal area in the pre-acquired narrow band endoscopic image can be segmented by a pre-trained abnormal area segmentation model to obtain an abnormal area segmentation map and an area of the target abnormal area. The abnormal area segmentation model is preferably Unet++, and the label thereof can be outlined by a professional nasopharyngeal laryngoscope physician on the boundary of the throat lesion area.

[0116] 202. Determine an area ratio feature of the target abnormal area based on the area of the target abnormal area and the size information of the narrow band endoscopic image.

[0117] The calculation formula of the area ratio feature label1 of the target abnormal area is as follows:

[0118]

[0119] S1 is the area of the target abnormal area, and W and H are the width and height of the narrow band endoscopic image, respectively.

[0120] 203. Segment the microvessel area in the abnormal area segmentation map to obtain a microvessel segmentation map.

[0121] In the embodiments of the present application, the microvessel area in the abnormal area segmentation map can be segmented by a pre-trained blood vessel segmentation model to obtain a microvessel segmentation map. The blood vessel segmentation model is preferably Unet++, and the label thereof can be outlined by a professional nasopharyngeal laryngoscope physician on the blood vessel boundary.

[0122] 204. Obtain a first number ratio feature of abnormal microvessels in the microvessel segmentation map and a second number ratio feature of hollow microvessels in the abnormal microvessels.

[0123] In some embodiments of the present application, the first number proportion feature of the abnormal microvessels in the microvessel segmentation map is obtained by steps A1 to A3.

[0124] A1, extract each microvessel in the microvessel segmentation map to obtain a first microvessel set, and count a first total number of microvessels in the first microvessel set.

[0125] Wherein, the first microvessel set includes all microvessels in the microvessel segmentation map.

[0126] In some embodiments of the present application, the extraction of each microvessel in the microvessel segmentation map to obtain a first microvessel set includes steps B1 to B3:

[0127] B1, on the basis of connected domain, obtain the minimum circumscribed horizontal rectangular frame of each microvessel in the microvessel segmentation map.

[0128] Wherein, the present application can adopt the connected domain method to obtain the minimum circumscribed horizontal rectangular frame of each microvessel in the microvessel segmentation map.

[0129] B2, and each connected domain is segmented from the microvessel segmentation map to obtain a plurality of sub-connected domains.

[0130] B3, based on the minimum circumscribed horizontal rectangular frame and the plurality of sub-connected domains, each microvessel in the microvessel segmentation map is obtained to obtain a first microvessel set.

[0131] In the embodiments of the present application, if there is only one connected domain in the rectangular frame, it is the target microvessel (one root) at this time, if there are multiple connected domains in the rectangular frame, the 4 boundary method is used for judgment: check whether there are pixels in all connected domains in the rectangular frame that appear on the 4 boundary of the rectangular frame at the same time, if less than 4 boundaries, the connected domain is excluded; finally, the single blood vessel is as follows Figure 3 As shown, A and B are single blood vessels.

[0132] A2, obtain the abnormal microvessels in the first microvessel set to obtain a second microvessel set, and count a second total number of abnormal microvessels in the second microvessel set.

[0133] Wherein, the abnormal blood vessel is a blood vessel with abnormal shape, size, etc., and the following mainly takes the diameter of the blood vessel as an example to illustrate how to obtain the first number proportion feature of the abnormal microvessels in the microvessel segmentation map.

[0134] In some embodiments of the present application, the abnormal microvessels in the first microvessel set are obtained by steps C1 to C6:

[0135] C1, obtaining a first diameter list set of each microvessel in the first microvessel set.

[0136] The first diameter list set includes a plurality of diameters of each microvessel, and the plurality of diameters are arranged in a preset size order.

[0137] In one embodiment, the first diameter list set is d_list = [d0, d1... dm].

[0138] C2, calculating a first diameter mean value of the first diameter list set of each microvessel.

[0139] In one embodiment, the first diameter mean value of the first diameter list set of each microvessel is d_mean:

[0140] Wherein d_mean = mean(d_list).

[0141] C3, based on the first diameter mean value, filtering out a second diameter list set greater than the first diameter mean value and a third diameter list set less than the first diameter mean value from the first diameter list set of each microvessel.

[0142] In one embodiment, the second diameter list set d_max_list is obtained by averaging the diameters greater than the first diameter mean value d_mean from the first diameter list set d_list of each microvessel, and the third diameter list set d_min_list is obtained by averaging the diameters less than the first diameter mean value d_mean.

[0143] C4, respectively calculating a second diameter mean value of the second diameter list set and a third diameter mean value of the third diameter list set.

[0144] In one embodiment, the second diameter mean value d_mean_max of the second diameter list set d_max_list is calculated, and the third diameter mean value d_mean_min of the third diameter list set d_min_list is calculated.

[0145] C5, comparing the second diameter mean value and the third diameter mean value to obtain a second ratio.

[0146] In one embodiment, the second ratio ω2 is obtained by comparing the second diameter mean value and the third diameter mean value, i.e. ω2 = d_mean_max / d_mean_min.

[0147] C6, based on the second ratio and a preset ratio threshold, determining each microvessel to obtain an abnormal microvessel in the first microvessel set.

[0148] The preset ratio threshold value can be set according to actual conditions.

[0149] A3, comparing the second total value with the first total value to obtain a first ratio, and taking the first ratio as a first quantity proportion feature of the abnormal microvessel in the microvessel segmentation graph.

[0150] In a specific embodiment, the first total value is N, the second total value is n, and then the first ratio ω1 = n / N, and thus the first quantity proportion feature label2 = n / N is obtained.

[0151] In some embodiments of the present application, the second quantity proportion feature of the hollow microvessel in the abnormal microvessel is obtained, including steps D1 to D4:

[0152] D1, performing color inversion processing on the minimum circumscribed rectangle frame of the abnormal microvessel in the second microvessel set to obtain a first black and white inversion graph.

[0153] In the embodiments of the present application, the color inversion processing is to perform black and white color inversion on the minimum circumscribed rectangle frame of the abnormal microvessel in the second microvessel set, as shown in the following formula: Figure 4 Figure 4 The left vessel graphs of the upper and lower pairs are the vessel graphs before processing, and the right vessel graphs thereof are the vessel graphs after processing, so that the blood vessels can be more obvious, facilitating subsequent image processing.

[0154] D2, based on a preset connected domain method, traversing the first black and white inversion graph and counting the number of connected domains in the first black and white inversion graph.

[0155] D3, if the number of connected domains is greater than a preset connected domain threshold value, it is determined that the abnormal microvessel is a hollow microvessel, and a third total value of the hollow microvessel in the second microvessel set is counted.

[0156] The connected domain threshold value can be set according to actual needs, and the connected domain threshold value is preferably 2.

[0157] D4, comparing the third total value with the second total value to obtain a third ratio, and taking the third ratio as a second quantity proportion feature of the hollow microvessel in the abnormal microvessel.

[0158] In a specific embodiment, the second quantity proportion feature is label3 = n A / n, wherein n is the second total value, and n A is the third total value.

[0159] 205, obtaining a third quantity proportion feature of a shape-irregular microvessel in the microvessel segmentation graph.​

[0160] In some embodiments of the present application, the acquiring the third quantity proportion feature of the irregular shape microvessel in the microvessel segmentation map comprises steps E1 to E6:

[0161] E1, performing color reversal processing on the minimum circumscribed rectangle frame of each microvessel in the first microvessel cluster to obtain a second black and white reversed image.

[0162] The principle of color reversal processing in the embodiments of the present application is the same as the color reversal processing in step D1 described above, and specific reference can be made to the introduction of the above content, which will not be repeated here.

[0163] E2, along the width and height direction of the minimum circumscribed rectangle frame of each microvessel point, a preset length is a target interval, and a plurality of first straight lines and a plurality of second straight lines are made in the minimum circumscribed rectangle frame of each microvessel parallel to the width and height.

[0164] In the embodiments of the present application, as shown in the following Figure 5 It can be understood that a plurality of horizontal lines and a plurality of vertical lines are arranged in the minimum circumscribed rectangle frame of each microvessel, and the horizontal lines / vertical lines are arranged according to the preset interval, wherein the target interval Δl of the preset length can be set according to actual needs.

[0165] E3, respectively counting the first intersection number and the second intersection number of the plurality of first straight lines and the plurality of second straight lines with the target region boundary line in the minimum circumscribed rectangle frame of each microvessel.

[0166] Wherein, the target region boundary line is the boundary line of the blood vessel corresponding region, as shown in the following Figure 5 The target region boundary line is the boundary line of the black region and the white region, and the first intersection number w ni and the second intersection number h nj are counted.

[0167] E4, based on the first intersection number, the second intersection number, the target interval and the size information of the minimum circumscribed rectangle frame of each microvessel, calculating the irregular quantization value of each microvessel.

[0168] Specifically, the calculation method of the irregular quantization value nul of each microvessel is as follows:

[0169]

[0170] Wherein, W1 and H1 are the width and height in the size information of the minimum circumscribed rectangle frame of each microvessel, Δl is the length of the target interval, w ni is the first intersection number and h nj is the second intersection number.

[0171] E5, judging shape irregular microvessels in the first microvessel set based on the irregular quantification value of each microvessel and a preset irregular quantification threshold value, and counting a fourth total number value of the shape irregular microvessels in the first microvessel set.

[0172] The irregular quantification threshold value can be set according to actual needs, which is not limited herein. Specifically, when the irregular quantification value of the microvessel is greater than the irregular quantification threshold value, the microvessel is determined to be a shape irregular microvessel, and the fourth total number value k of the shape irregular microvessels in the first microvessel set is counted.

[0173] E6, comparing the fourth total number value with the first total number value to obtain a fourth ratio value, and taking the fourth ratio value as a third quantity proportion feature of the shape irregular microvessels in the microvessel segmentation map.

[0174] The calculation method of the third quantity proportion feature label4 is as follows:

[0175] label4=k / N;

[0176] Wherein, k is the fourth total number value, and N is the first total number value.

[0177] 206, obtaining an abnormal microvessel aggregation degree feature, a microvessel density feature and an abnormal microvessel color quantification feature in the target abnormal area.

[0178] In the embodiments of the present application, the abnormal microvessel aggregation degree feature in the target abnormal area is obtained, including steps F1 to F6:

[0179] F1, obtaining the center of each microvessel (x i ,y i ) and the blood vessel area area i by the connected domain method;

[0180] F2, weighted calculation of the equivalent center coordinates of each microvessel, the calculation formula is:

[0181]

[0182] F3, calculating the equivalent center O(x0,y0) of the target abnormal area;

[0183] Specifically, the principle of calculating the equivalent center is as described in step F2, which is not repeated here.

[0184] F4, calculating the equivalent center P1(x1,y1) of the abnormal microvessel in the target abnormal area;

[0185] Specifically, the principle of calculating the equivalent centroid is as described in step F2, which will not be repeated here.

[0186] F5, calculating the equivalent centroid P2(x2, y2) of the normal microvessel in the target abnormal area;

[0187] Specifically, the principle of calculating the equivalent centroid is as described in step F2, which will not be repeated here.

[0188] F6, based on the equivalent centroid of the target abnormal area, the equivalent centroid of the abnormal microvessel in the target abnormal area, and the equivalent centroid of the normal microvessel in the target abnormal area, determining the abnormal microvessel aggregation degree feature in the target abnormal area.

[0189] Wherein, the calculation method of the abnormal microvessel aggregation degree feature label5 in the target abnormal area is as follows:

[0190]

[0191] In the embodiments of the present application, all the blood vessel areas and S2 of the target abnormal area can be obtained on the basis of the connected domain, and then the microvessel density feature label6 in the target abnormal area is calculated:

[0192] label6=S2 / S1, wherein S1 is the area of the target abnormal area, and S2 is the sum of all the blood vessel areas in the target abnormal area.

[0193] In the embodiments of the present application, the abnormal microvessel color quantification feature in the target abnormal area is obtained, including steps G1 to G3:

[0194] G1, calculating the color mean value of each blood vessel in the second microvessel concentration to obtain the color mean value list Ycolor mean _list, and then calculating the first variance std(Ycolor mean_list );

[0195] Specifically, the calculation process is as follows:

[0196] H1, obtaining all the color feature sets of the abnormal microvessel in the target abnormal area by the getcolors() method provided by PIL, such as color=[(r1, g1, b1), (r2, g2, b2)…(r n ,g n ,b n )].

[0197] H2, eliminating (r j ,g j ,b jelements whose values are all 0;

[0198] H3, calculating the color feature list mean of the remaining elements;

[0199]

[0200] G2, calculating the color mean of each blood vessel in the blood vessel set of normal blood vessels, obtaining the inter-normal blood vessel color mean list Zcolor mean _list, and then calculating the second variance thereof;

[0201] G3, based on the first variance and the second variance, calculating the abnormal microvessel color quantification feature label7 in the target abnormal area:

[0202]

[0203] 207, adding the area ratio feature, the first number ratio feature, the second number ratio feature, the third number ratio feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantification feature to the preset laryngeal image feature set.

[0204] The laryngeal image feature set is used to determine the abnormality degree of the larynx of the patient, which can include laryngeal injury abnormality, polyp abnormality, and tumor cancer abnormality.

[0205] The present application is exemplified by tumor cancer abnormality.

[0206] In some embodiments of the present application, the adding of the area ratio feature, the first number ratio feature, the second number ratio feature, the third number ratio feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantification feature to the preset laryngeal image feature set comprises steps I1 to I3:

[0207] I1, adding the area ratio feature, the first number ratio feature, the second number ratio feature, the third number ratio feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantification feature to the preset laryngeal image feature set.

[0208] I2, weighting and fitting all features in the laryngeal image feature set to obtain the abnormality degree coefficient θ of the larynx.

[0209] I3, determining the abnormality degree of the larynx of the patient based on the abnormality degree coefficient and the preset abnormality degree threshold τ.

[0210] The abnormality degree threshold τ can be set according to actual needs, which is not limited herein.

[0211] Specifically, the application illustrates the abnormal degree of tumor cancer in the larynx. If θ≤τ, it is determined that the abnormal degree of the larynx of the patient is non-in situ cancer; if θ>τ, it is determined that the abnormal degree of the larynx of the patient is in situ cancer.

[0212] The application provides a feature determination method of a larynx image. Compared with a traditional method, the application creatively proposes, in a case that the abnormal degree of the larynx of a patient cannot be determined efficiently and accurately, segmenting a target abnormal area of a narrow band endoscope image of the larynx of the patient, then comprehensively analyzing the target abnormal area to obtain a plurality of feature indexes corresponding to the target abnormal area, and finally comprehensively analyzing the plurality of feature indexes, thereby efficiently and accurately determining the abnormal degree of the larynx of the patient.

[0213] In order to better implement the feature determination method of the larynx image in the embodiment of the application, on the basis of the feature determination method of the larynx image, the embodiment of the application further provides a feature determination device of a larynx image, as shown in Figure 6 The feature determination device 600 of the larynx image comprises:

[0214] A first segmentation unit 601 is configured to segment a target abnormal area in a narrow band endoscope image pre-acquired, to obtain an abnormal area segmentation map and an area of the target abnormal area, the narrow band endoscope image being a narrow band endoscope image for photographing a larynx of a patient;

[0215] A first determination unit 602 is configured to determine an area proportion feature of the target abnormal area based on the area of the target abnormal area and size information of the narrow band endoscope image;

[0216] A second segmentation unit 603 is configured to segment a microvessel area in the abnormal area segmentation map, to obtain a microvessel segmentation map;

[0217] A first acquisition unit 604 is configured to acquire a first number proportion feature of abnormal microvessels in the microvessel segmentation map and a second number proportion feature of hollow microvessels in the abnormal microvessels;

[0218] A second acquisition unit 605 is configured to acquire a third number proportion feature of irregular shape microvessels in the microvessel segmentation map;

[0219] A third acquisition unit 606 is configured to acquire an abnormal microvessel aggregation degree feature, a microvessel density feature, and an abnormal microvessel color quantization feature in the target abnormal area;

[0220] The first adding unit 607 is configured to add the area proportion feature, the first quantity proportion feature, the second quantity proportion feature, the third quantity proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantization feature to a preset laryngeal image feature set, and the laryngeal image feature set is used to determine the abnormal degree of the larynx of the patient.

[0221] In some embodiments of the present application, the first obtaining unit 604 specifically comprises:

[0222] The first extracting unit is configured to extract each microvessel in the microvessel segmentation map to obtain a first microvessel set, and count a first total number of microvessels in the first microvessel set.

[0223] The second obtaining unit 605 is configured to obtain abnormal microvessels in the first microvessel set to obtain a second microvessel set, and count a second total number of abnormal microvessels in the second microvessel set.

[0224] The third obtaining unit 606 is configured to compare the second total number with the first total number to obtain a first ratio, and take the first ratio as a first quantity proportion feature of abnormal microvessels in the microvessel segmentation map.

[0225] In some embodiments of the present application, the first extracting unit is specifically configured to:

[0226] On the basis of the connected domain, a minimum circumscribed horizontal rectangular frame of each microvessel in the microvessel segmentation map is obtained.

[0227] Each connected domain in the microvessel segmentation map is segmented to obtain a plurality of sub-connected domains.

[0228] Based on the minimum circumscribed horizontal rectangular frame and the plurality of sub-connected domains, each microvessel in the microvessel segmentation map is obtained to obtain a first microvessel set.

[0229] In some embodiments of the present application, the second obtaining unit 605 is specifically configured to:

[0230] A first diameter list set of each microvessel in the first microvessel set is obtained, and the first diameter list set includes a plurality of diameters of each microvessel, and the plurality of diameters are arranged in a preset size order.

[0231] A first diameter mean value of the first diameter list set of each microvessel is calculated.

[0232] Based on the first diameter mean value, a second diameter list set greater than the first diameter mean value and a third diameter list set smaller than the first diameter mean value are screened from the first diameter list set of each microvessel.

[0233] respectively calculate a second diameter mean value of the second diameter list set and a third diameter mean value of the third diameter list set;

[0234] compare the second diameter mean value and the third diameter mean value to obtain a second ratio value;

[0235] based on the second ratio value and a preset ratio threshold value, determine each microvessel in the first microvessel set to obtain an abnormal microvessel in the first microvessel set.

[0236] In some embodiments of the present application, the first acquisition unit 604 is further configured to:

[0237] perform color inversion processing on the minimum circumscribed rectangular frame of the abnormal microvessel in the second microvessel set to obtain a first black-and-white inversion image;

[0238] based on a preset connected domain method, traverse the first black-and-white inversion image and count a number of connected domains in the first black-and-white inversion image;

[0239] if the number of connected domains is greater than a preset connected domain threshold value, determine that the abnormal microvessel is a hollow microvessel, and count a third total value of hollow microvessels in the second microvessel set;

[0240] compare the third total value with the second total value to obtain a third ratio value, and take the third ratio value as a second quantity ratio feature of the hollow microvessel in the abnormal microvessel.

[0241] In some embodiments of the present application, the second acquisition unit 605 is specifically configured to:

[0242] perform color inversion processing on the minimum circumscribed rectangular frame of each microvessel in the first microvessel set to obtain a second black-and-white inversion image;

[0243] along the width and height of the minimum circumscribed rectangular frame of each microvessel point, preset a length as a target interval, and make a plurality of first straight lines and a plurality of second straight lines parallel to the width and height in the minimum circumscribed rectangular frame of each microvessel;

[0244] respectively count a first intersection number and a second intersection number of the plurality of first straight lines and the plurality of second straight lines with the target region boundary line in the minimum circumscribed rectangular frame of each microvessel;

[0245] based on the first intersection number, the second intersection number, the target interval, and size information of the minimum circumscribed rectangular frame of each microvessel, calculate an irregular quantization value of each microvessel;

[0246] determine shape irregular microvessels in the first microvessel set based on the irregular quantification value of each microvessel and a preset irregular quantification threshold value, and count a fourth total number value of the shape irregular microvessels in the first microvessel set;

[0247] compare the fourth total number value with the first total number value to obtain a fourth ratio value, and use the fourth ratio value as a third quantity proportion feature of the shape irregular microvessels in the microvessel segmentation map.

[0248] In some embodiments of the present application, the first adding unit 607 is specifically configured to:

[0249] add the area proportion feature, the first quantity proportion feature, the second quantity proportion feature, the third quantity proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantification feature to a preset laryngeal image feature set;

[0250] weight and fit all features in the laryngeal image feature set to obtain an abnormal degree coefficient of the larynx;

[0251] determine the abnormal degree of the larynx of the patient based on the abnormal degree coefficient and a preset abnormal degree threshold value.

[0252] The throat image feature determination device provided in the application comprises a first segmentation unit 601, which is configured to segment a target abnormal area in a narrow-band endoscope image obtained in advance, to obtain an abnormal area segmentation map and an area of the target abnormal area, the narrow-band endoscope image being a narrow-band endoscope image taken for a throat of a patient; a first determination unit 602, which is configured to determine an area proportion feature of the target abnormal area based on the area of the target abnormal area and size information of the narrow-band endoscope image; a second segmentation unit 603, which is configured to segment a microvessel area in the abnormal area segmentation map, to obtain a microvessel segmentation map; a first acquisition unit 604, which is configured to acquire a first number proportion feature of abnormal microvessels in the microvessel segmentation map, a second number proportion feature of hollow microvessels in the abnormal microvessels; a second acquisition unit 605, which is configured to acquire a third number proportion feature of irregular-shape microvessels in the microvessel segmentation map; a third acquisition unit 606, which is configured to acquire an abnormal microvessel aggregation degree feature, a microvessel density feature and an abnormal microvessel color quantization feature in the target abnormal area; and a first adding unit 607, which is configured to add the area proportion feature, the first number proportion feature, the second number proportion feature, the third number proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantization feature to a preset throat image feature set, the throat image feature set being used to determine an abnormal degree of the throat of the patient. Compared with the traditional device, in the case that the abnormal degree of the throat of the patient cannot be determined efficiently and accurately, the application creatively proposes that the target abnormal area of the narrow-band endoscope image of the throat of the patient is segmented, then the target abnormal area is comprehensively analyzed to obtain a plurality of corresponding feature indexes, and finally the plurality of feature indexes are comprehensively analyzed, so that the abnormal degree of the throat of the patient is determined efficiently and accurately.

[0253] In addition to the above-mentioned throat image feature determination method and device, the embodiment of the application further provides a computer device integrated with any of the throat image feature determination devices provided in the embodiments of the application. The computer device comprises:

[0254] one or more processors;

[0255] a memory; and

[0256] one or more application programs, wherein the one or more application programs are stored in the memory and are configured to perform the operations of any of the methods in any of the throat image feature determination method embodiments by the processor.

[0257] The embodiment of the application further provides a computer device integrated with any of the throat image feature determination devices provided in the embodiments of the application. As Figure 7As shown in the figure, it shows a structural schematic diagram of a computer device related to embodiments of the present application, in particular:

[0258] The computer device can include a processor 701 with one or more processing cores, a storage unit 702 with one or more computer readable storage media, a power supply 703, and an input unit 704, etc. Those skilled in the art can understand that, Figure 7 The computer device structure shown in the figure does not constitute a limitation on the computer device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements. Among them:

[0259] The processor 701 is the control center of the computer device, which connects various parts of the computer device through various interfaces and lines, executes various functions and processes data of the computer device by running or executing software programs and / or modules stored in the storage unit 702, and calling data stored in the storage unit 702, thereby overall monitoring the computer device. Optionally, the processor 701 can include one or more processing cores; preferably, the processor 701 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application program, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 701.

[0260] The storage unit 702 can be used to store software programs and modules, and the processor 701 executes various functions and data processing by running the software programs and modules stored in the storage unit 702. The storage unit 702 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), etc.; the data storage area can store data created according to the use of the computer device, etc. In addition, the storage unit 702 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device. Accordingly, the storage unit 702 can also include a memory controller to provide access for the processor 701 to the storage unit 702.

[0261] The computer device also includes a power supply 703 for powering various components, and preferably the power supply 703 can be logically connected to the processor 701 through a power management system, so as to realize functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 703 can also include one or more direct current or alternating current power supplies, a recharging system, a power supply failure detection circuit, a power supply converter or inverter, a power supply state indicator, etc. Any component.

[0262] The computer device can also include an input unit 704, which can be used to receive inputted digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.

[0263] Although not shown, the computer device can also include a display unit and the like, which will not be described here. Specifically in the embodiments of the present application, the processor 701 in the computer device will load the executable file corresponding to the process of one or more application programs into the storage unit 702 according to the following instructions, and run the application program stored in the storage unit 702 by the processor 701, thereby realizing various functions, such as the above steps 201 to 207:

[0264] The present application provides a laryngeal image feature determination method. Compared with the traditional method, in the case that the abnormality degree of the larynx of a patient cannot be determined efficiently and accurately, the present application creatively proposes that the target abnormal area of the narrow band endoscopic image of the larynx of the patient is segmented, then the target abnormal area is comprehensively analyzed to obtain a plurality of corresponding feature indexes, and finally the plurality of feature indexes are comprehensively analyzed, thereby efficiently and accurately determining the abnormality degree of the larynx of the patient.

[0265] To this end, the embodiments of the present application provide a computer readable storage medium, which can include a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. The computer readable storage medium stores a plurality of instructions, which can be loaded by a processor to execute the steps in any of the laryngeal image feature determination methods provided by the embodiments of the present application. For example, the instructions can execute the above steps 201 to 207.

[0266] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0267] The above describes in detail the laryngeal image feature determination method, device and related equipment provided by the embodiments of the present application. The specific examples are applied to explain the principles and implementation modes of the present application. The above embodiment descriptions are only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed; in view of the above, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method of feature determination of a laryngeal image, characterized by, The method comprises: segmenting a target abnormal area in a pre-acquired narrow band endoscopic image to obtain an abnormal area segmentation map and an area of the target abnormal area, the narrow band endoscopic image being a narrow band endoscopic image taken for a throat of a patient; determining an area proportion feature of the target abnormal area based on the area of the target abnormal area and size information of the narrow band endoscopic image; segmenting a microvessel area in the abnormal area segmentation map to obtain a microvessel segmentation map; obtaining a first number proportion feature of abnormal microvessels in the microvessel segmentation map, a second number proportion feature of hollow microvessels in the abnormal microvessels; obtaining a third number proportion feature of irregular shape microvessels in the microvessel segmentation map; obtaining an abnormal microvessel aggregation degree feature, a microvessel density feature, and an abnormal microvessel color quantization feature in the target abnormal area; adding the area proportion feature, the first number proportion feature, the second number proportion feature, the third number proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantization feature to a preset throat image feature set, the throat image feature set being used to determine an abnormality degree of the throat of the patient.

2. The method of claim 1, wherein The method comprises: extracting each microvessel in the microvessel segmentation map to obtain a first microvessel set, and counting a first total number of microvessels in the first microvessel set; obtaining abnormal microvessels in the first microvessel set to obtain a second microvessel set, and counting a second total number of abnormal microvessels in the second microvessel set; comparing the second total number with the first total number to obtain a first ratio, and taking the first ratio as the first number proportion feature of the abnormal microvessels in the microvessel segmentation map.

3. The method of claim 2, wherein The method comprises: obtaining a minimum circumscribed horizontal rectangular frame of each microvessel in the microvessel segmentation map based on a connected domain; and segmenting each connected domain from the microvessel segmentation map to obtain a plurality of sub-connected domains; based on the minimum circumscribed horizontal rectangular frame and the plurality of sub-connected domains, obtaining each microvessel in the microvessel segmentation map to obtain a first microvessel set.

4. The method of claim 2, wherein The method comprises: obtaining a first diameter list set of each microvessel in the first microvessel set, the first diameter list set comprising a plurality of diameters of each microvessel, the plurality of diameters being arranged in a preset size order; calculating a first diameter mean value of the first diameter list set of each microvessel; based on the first diameter mean value, filtering a second diameter list set greater than the first diameter mean value and a third diameter list set smaller than the first diameter mean value from the first diameter list set of each microvessel; respectively calculating a second diameter mean value of the second diameter list set and a third diameter mean value of the third diameter list set; comparing the second diameter mean value with the third diameter mean value to obtain a second ratio; Based on the second ratio and a preset ratio threshold, each microvessel is determined to obtain an abnormal microvessel in the first microvessel set.

5. The method of claim 2, wherein The second number proportion feature of the hollow microvessel in the abnormal microvessel comprises: The minimum circumscribed rectangle frame of the abnormal microvessel in the second microvessel set is processed by color inversion to obtain a first black and white inversion image; Based on a preset connected domain method, the first black and white inversion image is traversed, and the number of connected domains in the first black and white inversion image is counted; If the number of connected domains is greater than a preset connected domain threshold, it is determined that the abnormal microvessel is a hollow microvessel, and a third total value of the hollow microvessel in the second microvessel set is counted; The third total value is compared with the second total value to obtain a third ratio, and the third ratio is taken as the second number proportion feature of the hollow microvessel in the abnormal microvessel.

6. The method of claim 2, wherein The third number proportion feature of the irregular shape microvessel in the microvessel segmentation image comprises: The minimum circumscribed rectangle frame of each microvessel in the first microvessel set is processed by color inversion to obtain a second black and white inversion image; Along the width and height of the minimum circumscribed rectangle frame of each microvessel, a plurality of first straight lines and a plurality of second straight lines are drawn in the minimum circumscribed rectangle frame of each microvessel at a preset length as a target interval; The first intersection number and the second intersection number of the target region boundary line in the minimum circumscribed rectangle frame of each microvessel are counted respectively. Based on the first intersection number, the second intersection number, the target interval, and the size information of the minimum circumscribed rectangle frame of each microvessel, the irregular quantization value of each microvessel is calculated. Based on the irregular quantization value of each microvessel and a preset irregular quantization threshold, the irregular shape microvessel in the first microvessel set is determined, and a fourth total value of the irregular shape microvessel in the first microvessel set is counted. The fourth total value is compared with the first total value to obtain a fourth ratio, and the fourth ratio is taken as the third number proportion feature of the irregular shape microvessel in the microvessel segmentation image.

7. The method of claim 1, wherein The area proportion feature, the first number proportion feature, the second number proportion feature, the third number proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantization feature are added to a preset laryngeal image feature set, comprising: The area proportion feature, the first number proportion feature, the second number proportion feature, the third number proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature, and the abnormal microvessel color quantization feature are added to a preset laryngeal image feature set; All features in the laryngeal image feature set are weighted and fitted to obtain an abnormal degree coefficient of the larynx; Based on the abnormal degree coefficient and a preset abnormal degree threshold, the abnormal degree of the larynx of the patient is determined.

8. A laryngeal image feature determination apparatus characterized by comprising: The device comprises: The first segmentation unit is configured to segment a target abnormal area in a narrow band endoscope image to obtain an abnormal area segmentation map and an area of the target abnormal area, the narrow band endoscope image being a narrow band endoscope image taken for a throat of a patient; The first determination unit is configured to determine an area proportion feature of the target abnormal area based on the area of the target abnormal area and size information of the narrow band endoscope image; The second segmentation unit is configured to segment a microvessel area in the abnormal area segmentation map to obtain a microvessel segmentation map; The first acquisition unit is configured to acquire a first number proportion feature of abnormal microvessels in the microvessel segmentation map, a second number proportion feature of hollow microvessels in the abnormal microvessels; The second acquisition unit is configured to acquire a third number proportion feature of irregular shape microvessels in the microvessel segmentation map; The third acquisition unit is configured to acquire an abnormal microvessel aggregation degree feature, a microvessel density feature and an abnormal microvessel color quantization feature in the target abnormal area; The first adding unit is configured to add the area proportion feature, the first number proportion feature, the second number proportion feature, the third number proportion feature, the abnormal microvessel aggregation degree feature, the microvessel density feature and the abnormal microvessel color quantization feature to a preset throat image feature set, the throat image feature set being used to determine an abnormal degree of the throat of the patient.

9. A computer device, comprising: The computer device comprises: one or more processors; a memory; and one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the feature determination method of the throat image according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps in the feature determination method of the throat image according to any one of claims 1 to 7.

Images