Search

CN-122023424-A - Intelligent classification evaluation method for leprosy nerve injury based on high-frequency ultrasonic multi-parameter fusion

CN122023424ACN 122023424 ACN122023424 ACN 122023424ACN-122023424-A

Abstract

The invention relates to the technical field of classified evaluation of leprosy nerve injury, in particular to an intelligent classified evaluation method of leprosy nerve injury based on high-frequency ultrasonic multi-parameter fusion, which adopts high-frequency ultrasonic to acquire multi-frame nerve images, analyzes gray scale trend to construct a closed path, screening out abnormal distribution frames to generate a consistent sequence, extracting region coverage, brightness morphology and texture arrangement characteristics, generating change description by comparing characteristic differences, matching preset classification rules according to description types, judging damage levels, and obtaining a leprosy nerve damage intelligent classification evaluation result. According to the invention, the neural area is defined in a self-adaptive manner by utilizing the gray level change of the image, the experience dependence is reduced, the spatial homology and stability of feature comparison are ensured by screening the consistent image sequence, the single index interference is weakened based on the integral trend tissue change description, the consistency and the distinction degree of the evaluation result under the complex background are ensured, and the applicability of the grading conclusion in the dynamic observation scene is enhanced.

Inventors

  • Jin Caifei
  • WANG BEN

Assignees

  • 浙江省皮肤病医院(浙江省皮肤病防治研究所、浙江武康疗养院、浙江省性病预防控制中心)

Dates

Publication Date
20260512
Application Date
20260414

Claims (8)

  1. 1. The intelligent classified evaluation method for leprosy injury based on high-frequency ultrasonic multi-parameter fusion is characterized by comprising the following steps of: S1, acquiring multi-frame nerve transection images of a ulnar nerve region of a patient, selecting a fixed reference position as a starting point, scanning along a plurality of preset directions, recording gray scale value change characteristics, and analyzing an increase and decrease range and direction distribution trend to obtain a direction gray scale change result of the images; S2, determining the coordinates of gray scale turning points in each preset direction according to the increase and decrease range and the distribution trend of gray scale values in the preset direction in the direction gray scale change result of the image, connecting the coordinates to form a closed path, determining the spatial position and the direction arrangement of the closed path, and determining the area range of the nerve cross section to obtain the expression content of the continuous boundary of the nerve cross section; S3, comparing the direction and the position of the continuous closed path in the multi-frame nerve cross-sectional images according to the spatial distribution of the continuous closed path in each cross-sectional image in the nerve cross-sectional continuous boundary expression content, screening out inconsistent cross-sectional image frames, retaining consistent cross-sectional image frames, and generating boundary consistent image sequence expression content; And S4, identifying the range change, the brightness distribution and the texture direction of the area surrounded by the closed path based on the reserved transverse image frames in the boundary consistent image sequence expression content to form change description content, and obtaining an image change description result.
  2. 2. The intelligent classification evaluation method for leprosy nerve injury based on high-frequency ultrasonic multi-parameter fusion according to claim 1, wherein the direction gray level change result of the image comprises a gray level change amplitude, a gray level change trend category and a direction gray level distribution characteristic of each direction, the nerve transverse continuous boundary expression content comprises a boundary geometric form, a boundary space closed state and a boundary integral continuous characteristic, the boundary consistent image sequence expression content comprises an effective image frame set, a boundary consistency identifier and a sequence stability characteristic, and the image change description result comprises a region form change type, a brightness distribution change type and a texture structure change type.
  3. 3. The intelligent classified evaluation method of leprosy injury based on high-frequency ultrasonic multi-parameter fusion according to claim 1, wherein the step of S1 is characterized in that: S101, carrying out continuous angle switching scanning on the region where the nerve of the leprosy patient ruler is positioned by adopting high-frequency ultrasonic equipment, acquiring echo image information according to an angle switching sequence, completing time sequence archiving of image frames according to scanning angles and image frame numbers, marking corresponding image coordinate origins and pixel reference scales, and generating an image frame sequence index set; S102, based on each frame of image in the image frame sequence index set, selecting an image coordinate origin as a fixed reference position, generating a plurality of direction paths in an image area according to the set direction quantity, calibrating pixel positions in the paths along each direction path in sequence, tracking a gray level change sequence, constructing a gray level continuous change record of each path, and acquiring a direction path gray level record table; S103, calling gray level change sequence data of each path in the direction path gray level record table, carding the change trend of gray level values in different directions, marking gray level continuous change sections in the paths according to adjacent gray level difference directions, and arranging the spatial correspondence of the change sections in different directions to obtain a direction gray level change result of the image.
  4. 4. The intelligent classified evaluation method of leprosy injury based on high-frequency ultrasonic multi-parameter fusion according to claim 1, wherein the step of S2 is characterized in that: S201, according to the increasing and decreasing range and the distribution trend of each preset direction gray scale value in the direction gray scale change result of the image, locating the position of the gray scale value from ascending to descending or from descending to ascending in each preset direction, screening out pixel points with continuous mutation characteristics according to the position of a turning part of a symbol of an adjacent gray scale change section, recording corresponding image coordinate values, and generating a direction turning space coordinate set; s202, calling space position coordinates corresponding to all directions in the direction turning space coordinate set, sequentially connecting turning points in all directions according to a preset direction arrangement sequence, forming a closed curve surrounding a nerve region in an image through line segment connection, and calibrating direction connecting points at a closed segment of the closed curve to obtain a continuous closed path coordinate sequence; And S203, locking the area covered by the closed path according to the distribution range of all the closed curves in the continuous closed path coordinate sequence in the image coordinate system, establishing a corresponding relation between the position numbers of all the pixel points in the area and the path boundary, extracting the image pixel space area surrounded by the continuous closed path, and obtaining the expression content of the nerve transverse continuous boundary.
  5. 5. The intelligent classified evaluation method of leprosy injury based on high-frequency ultrasonic multi-parameter fusion according to claim 1, wherein the step of S3 is characterized in that: S301, extracting the direction arrangement sequence and the center position coordinate of each frame of closed path in an image coordinate system according to the continuous closed path coordinate corresponding to each frame of image in the nerve transection continuous boundary expression content, constructing a direction arrangement index list corresponding to all frames by taking a preset path reference direction as a reference, and acquiring a closed path direction arrangement parameter set; S302, calling the direction arrangement index and the central position coordinate of each image frame in the closed path direction arrangement parameter set, pairing the direction arrangement index and the central position coordinate in pairs according to the sequence of the image frames, judging the difference value of the direction indexes, comparing the difference value with the position coordinate, marking out the image frame numbers with inconsistent direction indexes or out-of-range position offset, and generating a space sequence screening number set; And S303, eliminating corresponding image frame numbers according to the space sequence screening number set, reserving image frames with consistent direction arrangement sequence and stable position distribution from the original image frame sequence, reconstructing a sequence index for the reserved frames, and establishing space stable expression of a continuous closed path boundary sequence to obtain the expression content of the boundary consistent image sequence.
  6. 6. The intelligent classified evaluation method of leprosy injury based on high-frequency ultrasonic multi-parameter fusion according to claim 1, wherein the step of S4 is characterized in that: S401, extracting an image pixel set of an area surrounded by a path based on a continuous closed path of each frame image in the expression content of the boundary consistent image sequence, counting the corresponding relation between the total number of coordinates of pixel points in the surrounded area and the image size, recording the ratio occupied by each frame area according to the image number, and sequencing to obtain a surrounded area coverage ratio sequence; s402, calling an inner area of a closed path corresponding to each frame image in the surrounding area coverage ratio sequence, reading all pixel gray scale values in the area, dividing sub-areas according to the pixel position and the row-column direction, comparing the distribution density of the gray scale values of each sub-area, classifying to form a brightness distribution group, and obtaining a pixel brightness distribution feature group; S403, extracting the arrangement direction of pixels between adjacent boundaries in a closed path area according to the dividing boundaries of the brightness distribution in each image frame in the pixel brightness distribution characteristic group, judging the consistency and the offset degree of the arrangement direction of textures in the adjacent frames according to the angle change of the arrangement direction in an image coordinate system, and obtaining an image change description result.
  7. 7. The intelligent classified evaluation method of leprosy injury based on high-frequency ultrasonic multi-parameter fusion according to claim 1, wherein the method further comprises the following steps: S5, classifying and sorting the change types according to the change description types corresponding to each transverse image frame in the image change description results, matching with preset nerve injury classification rules to form a nerve injury classification evaluation conclusion, and completing the association expression with the image frames to obtain a leprosy nerve injury intelligent classification evaluation result; the intelligent classified evaluation result of the leprosy nerve injury comprises nerve injury class, class corresponding feature combination and image frame evaluation association identification.
  8. 8. The intelligent classified evaluation method of leprosy injury based on high-frequency ultrasonic multi-parameter fusion according to claim 7, wherein the step of S5 is characterized in that: S501, according to the change description type of each cross-section image frame in the image change description result, taking three characteristics of coverage change, pixel brightness distribution form and texture arrangement direction corresponding to each frame image as classification basis, calling image frame sequence numbers and corresponding characteristic values to carry out combination classification, and generating a change characteristic classification matrix; S502, according to each group of feature combination types in the change feature classification matrix, calling a matching relation set in a preset nerve injury classification rule, judging the injury level number corresponding to each group of types item by item, constructing corresponding items of an image frame number and a classification number, and acquiring a nerve injury classification number list; And S503, calling a grading number corresponding to each image frame in the nerve injury grading number list, mapping the grading number to a corresponding frame position in an original transverse image sequence, integrating grading information of each frame according to the image sequence, and assigning each grading number to a corresponding image frame record to obtain an intelligent grading evaluation result of the leprosy nerve injury.

Description

Intelligent classification evaluation method for leprosy nerve injury based on high-frequency ultrasonic multi-parameter fusion Technical Field The invention relates to the technical field of classified evaluation of leprosy nerve injury, in particular to an intelligent classified evaluation method of leprosy nerve injury based on high-frequency ultrasonic multi-parameter fusion. Background The technical field of classified evaluation of leprosy nerve injury comprises an evaluation and analysis method for the damage degree of peripheral nerve structures and functions of leprosy patients, and the core content mainly relates to objective identification and quantitative evaluation for pathological changes such as inflammation, edema, denaturation, necrosis and the like caused by leprosy bacillus invading peripheral nerves through means of medical imaging, physiological function detection and the like, and the whole technical field covers the contents such as neuroelectrophysiology detection technology, high-resolution imaging technology, tissue mechanics parameter evaluation, nerve function quantification standard and the like, and is widely applied to a plurality of medical links such as neuropathological diagnosis, disease typing, rehabilitation monitoring, therapeutic intervention and the like, and has definite disease directivity and evaluation flow standardization characteristics. The method is characterized in that a plurality of image characteristic parameters including nerve cross-sectional area, echo intensity, morphological boundary and internal structural integrity are extracted through collecting high-frequency ultrasonic images of affected peripheral nerves of a patient, a standard reference value system is established by combining past clinical data, nerve injury classification is realized by adopting parameter fusion judgment rules according to the change rules of different characteristic parameters in various levels of injuries, the related technical matters comprise ultrasonic imaging data acquisition, characteristic index selection, index standardization processing, characteristic fusion classification methods and the like, the nerve states are classified mainly by setting specific numerical ranges and judgment conditions, and classification judgment is completed by taking multi-angle and multi-section parameters of nerve cross-section scanning as evaluation bases. In the prior art, a plurality of preset image parameters are used as main judgment basis, a single frame or a small number of tangential plane measurement results are relied on in practical application, when the scanning angle changes or the nerve morphology presents a local irregular state, the characteristic corresponding relation among different images is easy to deviate, so that parameter combination unbalance is caused, meanwhile, when the grade division is carried out in a fixed interval, the integral change trend of a nerve region in a continuous image is difficult to reflect, under the superposition scene of a plurality of pathological states, the parameter performance tends to be similar, the differentiation degree of the grading result is reduced, and the stability and consistency of an evaluation conclusion in follow-up comparison and clinical decision are influenced. Disclosure of Invention In order to solve the technical problems in the prior art, the embodiment of the invention provides an intelligent classification evaluation method for leprosy nerve injury based on high-frequency ultrasonic multi-parameter fusion. The technical scheme is as follows: the intelligent classified evaluation method for leprosy nerve injury based on high-frequency ultrasonic multi-parameter fusion comprises the following steps: S1, acquiring multi-frame nerve transection images of a ulnar nerve region of a patient, selecting a fixed reference position as a starting point, scanning along a plurality of preset directions, recording gray scale value change characteristics, and analyzing an increase and decrease range and direction distribution trend to obtain a direction gray scale change result of the images; S2, determining the coordinates of gray scale turning points in each preset direction according to the increase and decrease range and the distribution trend of gray scale values in the preset direction in the direction gray scale change result of the image, connecting the coordinates to form a closed path, determining the spatial position and the direction arrangement of the closed path, and determining the area range of the nerve cross section to obtain the expression content of the continuous boundary of the nerve cross section; S3, comparing the direction and the position of the continuous closed path in the multi-frame nerve cross-sectional images according to the spatial distribution of the continuous closed path in each cross-sectional image in the nerve cross-sectional continuous boundary expression content, screening out incon