Contrast agent-based blood perfusion analysis method and system for skin burns

Abstract

The application belongs to the technical field of burn wound analysis. A skin burn analysis method and system based on contrast agent blood perfusion are provided. First, the fluorescence sequence after injecting the contrast agent is obtained, and the pre-processing sequence is generated by region cutting and down-sampling. Then, the fluorescence intensity time curve of each pixel point is extracted, and the peak time is determined to construct the peak time histogram. By smoothing and denoising the histogram, extreme value detection and weak peak removal, the threshold for distinguishing normal and burn tissues is adaptively selected. The initial mask is generated using the threshold, and the target burn area mask is obtained through connected domain analysis and morphological post-processing. Finally, the pseudo-color map is generated combined with the peak time. The application fully utilizes the time dynamic characteristics of blood perfusion, does not need to rely on skin texture or a large amount of training data, can accurately segment the burn area, and intuitively presents the microcirculation disorder degree, which significantly improves the objectivity and accuracy of burn depth evaluation.

Description

Technical Field

[0001] This invention relates to the field of burn wound analysis technology, specifically to a method and system for skin burn analysis based on contrast agent blood perfusion. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Accurate assessment of burn depth is crucial for clinical treatment decisions, directly determining the timing of debridement, skin grafting regimen, and prognosis. Currently, clinical practice relies primarily on visual observation and palpation by physicians for initial assessment, but this method is highly subjective, especially in differentiating between deep second- and third-degree burns, where significant errors exist. With the development of optical imaging technology, techniques such as laser Doppler flow imaging (LDI) and indocyanine green (ICG) fluorescence angiography are increasingly being applied to burn assessment. Among these, near-infrared fluorescence imaging-based blood perfusion analysis, through intravenous injection of contrast agents and irradiation with excitation light of a specific wavelength, can capture the dynamic changes in subcutaneous microcirculation in real time. This technology utilizes the significant differences in hemodynamics between normal and necrotic tissues, theoretically enabling objective quantification of burn depth. Compared to traditional methods, fluorescence imaging offers advantages such as non-contact, high sensitivity, and repeatability, clearly presenting the filling status of skin microvessels and providing burn surgeons with abundant blood perfusion information, making it a highly promising auxiliary diagnostic tool in modern burn treatment.

[0004] However, existing burn analysis methods based on contrast agent blood perfusion still face many technical bottlenecks in practical applications. First, the raw fluorescence sequence data is massive and contains a lot of background noise. The lack of an efficient preprocessing mechanism leads to an excessive computational burden for subsequent analysis, making it difficult to meet the real-time needs of clinical practice. Second, when distinguishing between normal skin and burn tissue, existing algorithms mostly rely on fixed thresholds or simple statistical averages, ignoring individual patient differences and the complex distribution characteristics of blood perfusion time in burn areas. This results in insufficient segmentation accuracy in areas with blurred boundaries, easily leading to misjudgments or omissions. In addition, the lack of targeted connected component analysis and morphological optimization strategies for the generated initial burn area mask often results in isolated noise points or regional breaks in the final results, failing to form a continuous and complete burn lesion outline. Finally, existing systems are relatively simple in their visualization presentation, failing to deeply integrate and render key hemodynamic time parameters and spatial location information. This limits doctors' intuitive understanding of burn depth and extent, affecting the accuracy of diagnosis and the formulation of treatment plans. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a skin burn analysis method and system based on contrast agent blood perfusion. This method can fully integrate fluorescence intensity and temporal dynamic information during blood perfusion, thereby more accurately dividing burned and non-burned areas. By introducing time-based visualization technology, the hemodynamic characteristics are presented intuitively, providing strong support for the auxiliary judgment of burn depth.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for analyzing skin burns based on contrast agent blood perfusion.

[0008] A method for analyzing skin burns based on contrast agent blood perfusion includes the following steps:

[0009] Obtain the fluorescence sequence of the skin area to be analyzed. The fluorescence sequence contains fluorescence intensity data over time collected under excitation light source irradiation after injection of contrast agent.

[0010] The region of interest of the fluorescence sequence is cropped and downsampled to generate a preprocessed fluorescence sequence;

[0011] Based on the preprocessed fluorescence sequence, the fluorescence intensity-time curve of each pixel is extracted, and the peak fluorescence intensity time of each pixel is determined.

[0012] The peak fluorescence times of all pixels in the preprocessed fluorescence sequence were statistically analyzed, a peak fluorescence time histogram was constructed, and a threshold for distinguishing normal skin from burned skin was selected based on the peak fluorescence time histogram.

[0013] The pixels in the preprocessed fluorescence sequence are binarized and classified using a threshold to generate an initial burn area mask;

[0014] Connectivity analysis and morphological post-processing are performed on the initial burn area mask to generate the target burn area mask.

[0015] A pseudo-color image of the burn area is generated based on the mask of the target burn area and the peak time of fluorescence intensity.

[0016] In one implementation of the first aspect of the present invention, the step of constructing the fluorescence intensity peak time histogram further includes:

[0017] The fluorescence intensity peak time histogram is smoothed and denoised to generate a smoothed histogram curve.

[0018] Local extremum detection is performed based on the smoothed histogram curve to identify sets of peaks and valleys.

[0019] Calculate the significance coefficient of each peak in the peak set, remove weak peaks with significance coefficients lower than the preset minimum significance coefficient, and obtain the main peak set.

[0020] In one implementation of the first aspect of the present invention, in the step of selecting the threshold for distinguishing between normal skin and burned skin, if the set of main peaks contains two main peaks, the following operations are performed: calculate the relative valley depth between the two main peaks, determine whether the relative valley depth meets the preset acceptance threshold condition, and if the acceptance threshold condition is met, determine the time value corresponding to the valley point between the two main peaks as the threshold.

[0021] If the set of main peaks contains only one main peak or does not meet the criteria for determining two peaks, then perform the following operations: determine the nearest valleys to the left and right of the main peak as candidate valleys, compare the depths of the candidate valleys relative to the main peaks, select the time value corresponding to the candidate valley with the more significant depth as the threshold, if the region of interest is determined to be a large area of ​​burns, then directly retain the entire region as the initial burn area mask, if the region of interest is determined to be a large area of ​​normal skin, then terminate the process.

[0022] In one implementation of the first aspect of the present invention, the connected component analysis in the step of generating the target burn area mask specifically includes:

[0023] Calculate the area of ​​each connected component in the initial burn area mask and determine the maximum connected area;

[0024] The adaptive area threshold is calculated based on the maximum connected area and the preset minimum number of retained pixels.

[0025] Delete connected components whose area is smaller than the adaptive area threshold and which are determined to be invalid masks.

[0026] In one implementation of the first aspect of the present invention, the morphological post-processing in the step of generating the target burn area mask specifically includes:

[0027] Identify background connected components; if a background connected component is not connected to the image boundary, it is determined to be a valid mask hole.

[0028] When the area of ​​an effective mask hole is smaller than the preset small effective mask area, the effective mask hole is filled.

[0029] A morphological closing operation is performed on the filled mask to eliminate regional gaps and merge neighboring fragments, thus obtaining the target burn area mask.

[0030] In one implementation of the first aspect of the present invention, the step of generating a pseudo-color image of the burn area specifically includes:

[0031] Traverse every pixel within the mask range of the target burn area;

[0032] Read the peak fluorescence intensity time of each pixel in the fluorescence intensity-time curve;

[0033] The peak fluorescence intensity time is mapped to a specific pseudo-color value, and the pixels within the mask range of the target burn area are rendered and colored.

[0034] Secondly, the present invention provides a skin burn analysis system based on contrast agent blood perfusion.

[0035] A skin burn analysis system based on contrast agent blood perfusion includes:

[0036] The data acquisition unit is configured to acquire the fluorescence sequence of the skin region to be analyzed, the fluorescence sequence containing fluorescence intensity data over time acquired under excitation light source irradiation after injection of contrast agent;

[0037] The preprocessing unit is configured to: perform region-of-interest (ROI) cropping and downsampling on the fluorescence sequence to generate a preprocessed fluorescence sequence;

[0038] The curve analysis unit is configured to: extract the fluorescence intensity time curve of each pixel based on the preprocessed fluorescence sequence, and determine the fluorescence intensity peak time corresponding to each pixel;

[0039] The threshold selection unit is configured to: statistically analyze the peak fluorescence intensity time of all pixels in the preprocessed fluorescence sequence, construct a peak fluorescence intensity time histogram, and select a threshold to distinguish between normal skin and burned skin based on the peak fluorescence intensity time histogram;

[0040] The initial screening mask unit is configured to: perform binarization classification of pixels in the preprocessed fluorescence sequence using a threshold to generate an initial burn area mask;

[0041] The post-processing unit is configured to perform connected component analysis and morphological post-processing on the initial burn area mask to generate the target burn area mask.

[0042] The visualization unit is configured to generate a pseudo-color image of the burn area based on the target burn area mask and the peak time of fluorescence intensity.

[0043] Thirdly, the present invention provides a computer device, comprising: a processor and a computer-readable storage medium;

[0044] A processor, adapted to execute computer programs;

[0045] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the skin burn analysis method based on contrast agent blood perfusion according to the first aspect of the present invention.

[0046] Fourthly, the present invention provides a computer-readable storage medium storing a computer program adapted to be loaded by a processor and to execute the skin burn analysis method based on contrast agent blood perfusion of the first aspect of the present invention.

[0047] Fifthly, the present invention provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the skin burn analysis method based on contrast agent blood perfusion according to the first aspect of the present invention.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] To address the technical problems of existing technologies, such as excessive computational load when processing raw fluorescence sequences, insufficient boundary segmentation accuracy due to reliance on fixed thresholds, and lack of effective post-processing strategies, this invention achieves high-precision segmentation of burn areas by constructing a fluorescence intensity peak-time histogram and adaptively selecting a discrimination threshold. This method delves into the dynamic changes in fluorescence intensity over time during blood perfusion, moving beyond static image analysis at a single moment. Instead, it utilizes the significant difference in the peak contrast agent arrival time between normal and necrotic tissues for differentiation. By statistically analyzing the peak-time distribution of all pixels, the algorithm automatically adapts to individual differences among patients and variations in lighting conditions, avoiding the blindness of manually setting thresholds. Furthermore, by combining connected component analysis and morphological post-processing steps, isolated noise points in the initial mask are effectively removed and internal holes are filled, generating a continuous and complete target mask for the burn area. This series of compact processing steps not only significantly improves the segmentation robustness under complex backgrounds and fuzzy boundary conditions, but also ensures that the final generated burn area contour closely matches the actual lesion, providing clinicians with a reliable basis for quantitative analysis and fundamentally solving the problems of high misjudgment rate and inconsistent results of traditional methods.

[0050] This invention creatively proposes a time-dimensional pseudo-color visualization scheme, transforming abstract hemodynamic parameters into intuitive visual information, greatly assisting doctors in assessing burn depth. In traditional imaging techniques, doctors often face a large number of grayscale images or video streams, making it difficult to directly capture the specific location and severity of microcirculatory disturbances from continuous frame sequences. This scheme extracts the peak fluorescence intensity time corresponding to each pixel and maps this time value to a specific pseudo-color value, creating a clear color contrast between areas with delayed blood perfusion (typically deep burns) and areas with normal perfusion. This visualization method not only preserves the spatial location information of the burn area but also deeply integrates the temporal delay characteristics of tissue perfusion, making the key diagnostic indicator "time" "visible." Doctors do not need complex signal processing backgrounds; they only need to observe the color distribution of the pseudo-color image to quickly identify the core area of ​​microcirculatory failure and its spread. This intuitive presentation shortens diagnostic time, reduces diagnostic bias caused by differences in human experience, and provides valuable reference for developing treatment plans such as debridement and skin grafting, significantly improving the efficiency and accuracy of clinical diagnosis and treatment.

[0051] Unlike traditional image processing methods that rely on skin texture, color, or structural features, this invention performs functional analysis entirely based on contrast-excited fluorescence video or image sequences, offering significant advantages in terms of independence and low computational cost. Traditional visible light image-based analysis methods are highly susceptible to interference from wound exudate, eschar color, ambient lighting, and shooting angle, and often require training deep neural network models with a large number of labeled samples. This not only makes data acquisition difficult but also results in poor model generalization ability, making it difficult to adapt to changing clinical scenarios. This invention takes a different approach, directly using blood perfusion as a physiological functional indicator as a criterion, completely avoiding the influence of unstable appearance features. Since the core of the algorithm lies in the distribution characteristics of statistical peak times and simple morphological operations, there is no need to build complex feature classifiers or require massive training datasets for model iteration. This allows the method to run in real time on ordinary computing devices, significantly reducing the hardware computing power requirements. At the same time, the system does not require the collection of any additional data (such as three-dimensional structured light or multispectral data) besides conventional fluorescence sequences, simplifying the data acquisition process, reducing equipment costs, and making this technology easier to promote and popularize in primary healthcare institutions, thus possessing extremely high practical value.

[0052] This invention cleverly utilizes the fundamental differences in blood flow input and clearance dynamics between normal skin and burn tissue. It completes the entire process from data preprocessing to final diagnostic result generation using only the fluorescence sequences routinely acquired by the system, demonstrating extremely high system integration and ease of operation. In specific implementation, the algorithm effectively reduces data redundancy and improves processing speed by cropping and downsampling the region of interest in the fluorescence sequence while preserving key hemodynamic information. Subsequently, by tracking the fluorescence intensity time curve of each pixel and locking its peak time, the transmission delay characteristics of the contrast agent in the microvascular network are accurately captured. This feature extraction method based on "peak time" is more specific than simply comparing the maximum fluorescence intensity or integral area, and can more sensitively reflect the blockage of microcirculation. The entire analysis process requires no additional sensors, markers, or complex calibration steps and is fully compatible with existing indocyanine green (ICG) fluorescence angiography equipment. This "sampling and analysis on demand" model not only simplifies clinical procedures and reduces patient waiting time, but also improves the consistency and repeatability of analysis results, making burn assessment more standardized and objective, and effectively overcoming the subjectivity and arbitrariness of manual visual assessment.

[0053] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0054] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0055] Figure 1 A schematic diagram illustrating the principle of a skin burn analysis method based on contrast agent blood perfusion, provided as an exemplary embodiment of the present invention;

[0056] Figure 2 A schematic flowchart of a skin burn analysis method based on contrast agent blood perfusion, provided as an exemplary embodiment of the present invention;

[0057] Figure 3 A standard, complete fluorescence profile is provided for an exemplary embodiment of the present invention;

[0058] Figure 4 Peak time histogram under ideal conditions provided as an exemplary embodiment of the present invention;

[0059] Figure 5 A schematic diagram of a skin burn analysis system based on contrast agent blood perfusion, provided as an exemplary embodiment of the present invention;

[0060] Figure 6A schematic diagram of a computer device provided for an exemplary embodiment of the present invention. Detailed Implementation

[0061] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0062] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0063] After a contrast agent is injected into the bloodstream, it emits fluorescence under the illumination of an excitation light source. This fluorescence can be captured by high-speed photography or real-time video recording to obtain a dynamic blood flow image, thus assessing tissue blood perfusion. Based on this principle, this invention proposes a burn analysis method based on contrast agent blood perfusion, such as... Figure 1 As shown, the original fluorescence sequence is first cropped and downsampled to generate preprocessed data. Then, fluorescence intensity time-varying curves are analyzed, and a histogram is constructed based on this for threshold selection. In the histogram analysis stage, smoothing and denoising, local extremum detection, and removal of weak peaks are performed sequentially to screen for the main peak. The classification criteria are then determined through bimodal peak determination and valley threshold logic. If a bimodal peak is identified, the process proceeds to mask generation and post-processing, specifically including effective mask identification, conditional filling, and morphological closing operations. Regardless of whether the post-processing branch is executed, the results are ultimately used for pseudo-color visualization, thus completing a full analytical loop from data processing to intuitive presentation.

[0064] More specifically, such as Figure 2 As shown, the process includes the following:

[0065] S201: Region of Interest (ROI) clipping and downsampling.

[0066] Removing non-interesting regions can significantly reduce computational load. Fluorescence intensity typically doesn't change abruptly in data sequences during practice; therefore, downsampling the data can also reduce computational load, thereby improving data processing speed.

[0067] S202: Pixel fluorescence intensity time curve analysis.

[0068] For a standard complete fluorescence curve, the curve is as follows: Figure 3 As shown, where, The peak fluorescence intensity time is the threshold value. Burns cause local and systemic increases in microvascular permeability, altered inflammatory blood flow, and edema formation. These physiological changes determine the input and removal kinetics of contrast agents within tissues, thus affecting the shape of the fluorescence time curve. Damaged skin exhibits the following significant changes in the curve representation compared to normal skin: a slower rise phase; a longer time to reach fluorescence intensity; and a slower decline after reaching the peak fluorescence intensity. Therefore, in actual testing, using the peak fluorescence intensity time as a threshold can more completely distinguish between skin areas and burn areas.

[0069] S203: Histogram analysis of fluorescence intensity peak time and threshold selection.

[0070] Analyze the time curve corresponding to each pixel in the fluorescence sequence after cropping and downsampling. The fluorescence intensity peak-time histogram requires statistically analyzing the peak time of each pixel. The peak time can be determined by a simple left-to-right maximum value search. The set of pixels corresponding to the largest times is then selected as the chosen burn area; this requires designing a reasonable threshold.

[0071] Assuming both normal and burned skin are present, the peak time histogram will exhibit a bimodal distribution, such as... Figure 4 As shown, since the time to reach peak intensity in normal skin is much shorter than that in burned skin, the peak on the left and its surrounding area can be considered normal, while the peak on the right and its surrounding area are burned areas. Therefore, the appropriate threshold selection is essential for distinguishing between burned and unburned areas.

[0072] The steps for selecting the threshold are as follows:

[0073] S203-1: Smoothing and noise reduction of histogram curves.

[0074] To reduce the impact of data noise, the curve needs to be smoothed and denoised:

[0075] (1);

[0076] (2);

[0077] This expression is a one-dimensional Gaussian convolution smoother, where, For the smoothed curve at position The value at; This is the Gaussian smoothing kernel function; The width of the Gaussian kernel function; To control the noise removal coefficient; This is the convolution operator; For the original curve (unsmoothed histogram) at position The value at; The Gaussian smoothing kernel function at the offset The value at; Represents the window length or the total length of the data sequence involved in the convolution calculation; This represents the coefficient used to control noise removal. This represents the value of the Gaussian smoothing kernel function; The independent variable represents the Gaussian kernel function.

[0078] S203-2: Local extremum detection.

[0079] Define the discrete difference as:

[0080] (3);

[0081] in, To smooth the curve at position The value at; To smooth the curve at position Discrepancy at point; To smooth the curve at position The value at that location.

[0082] Peak set (Local maximum) is represented as:

[0083] (4);

[0084] in, For the peak point The previous position ( Discrepancy at point ( ); Representative at the peak The discrepancy at the current position.

[0085] Valley Point Set (Local minimum) is represented as:

[0086] (5);

[0087] in, Represents the valley point The previous position ( Discrepancy at point ( ); Represents the valley point The discrepancy at the current position.

[0088] S203-3: Remove weak peaks.

[0089] Define the leftward traversal process: from the peak Start searching to the left along the curve until you encounter the first peak with a height greater than the summit. The point with the smallest ordinate value is defined as the point visited along this search path. Define the rightward traversal process: from peak Start searching to the right along the curve until you encounter the first peak with a height greater than the peak. The point with the smallest ordinate value is defined as the point visited along this search path. Then the expression for peak significance is:

[0090] (6);

[0091] in, For Peak The significance coefficient of is denoted as the lowest significance coefficient. Ignore all values ​​less than The significant sequence is used to remove weak peaks; This represents the peak point of the smoothed curve. The ordinate value at; The higher of the two reference valleys on the left and right sides is taken as the "local base" height of the peak.

[0092] S203-4: Selecting the Main Peak .

[0093] Take in Chinese Press If the top two peaks in descending order have only one significant peak, they should be treated as single peaks.

[0094] S203-5: Bimodal determination and valley threshold, taking the valley between the two peaks.

[0095] If there are two main peaks ,but:

[0096] (7);

[0097] in, Representatives at the two main peaks and Within the interval between, make the curve smooth. The x-coordinate of the point where the minimum value is obtained (i.e., the valley bottom position).

[0098] Define valley depth criteria:

[0099] (8);

[0100] (9);

[0101] in, The arithmetic mean of the heights of the two main peaks; Representing the left main peak The original histogram height at that location; Representing the right main peak The original histogram height at that location; Represents relative valley depth, indicating the "sinking depth" of the valley floor relative to the average height of the two peaks.

[0102] Acceptance threshold conditions:

[0103] (10);

[0104] in, This is a preset scaling factor used to indicate how much lower the valley must be than the average peak height to be considered "significant." Very small percentages (e.g., <5%) indicate an insignificant valley and are therefore unreliable.

[0105] S203-6: Lateral threshold selection for single peaks or weak secondary peaks.

[0106] When the bimodal condition is not met (secondary peak is too small or valley is shallow), the position of the primary peak is used. Decision-making, defining the nearest left and right valleys:

[0107] (11);

[0108] (12);

[0109] in, Main peak The x-coordinate of the nearest valley point on the left; Main peak The x-coordinate of the nearest valley point on the right.

[0110] Then normalize the peak:

[0111] (13);

[0112] in, Main peak The normalized position on the entire peak time axis, with a value range of [0,1]; This represents the minimum value of the x-axis of the entire histogram; This represents the maximum value of the x-axis of the entire histogram.

[0113] The left and right depths are:

[0114] (14);

[0115] (15);

[0116] in, Representative from the main peak To the nearest valley on the left The "depth difference"; Represents the peak point of the original histogram The height of the location; This represents the smoothed curve at the valley point on the left. The height of the location; Representative from the main peak To the nearest valley on the right The "depth difference"; This represents the valley point on the right side of the smoothed curve. The height of the location.

[0117] definition To allow for a depth comparison tolerance (comparing the relative depths of two candidate valleys to the left and right of the main peak to determine which side's valley is more significant as the final threshold), then:

[0118] like (The main peak is slightly to the right) and The region of interest was considered to be a large area of ​​burns, which was preserved entirely and could be directly visualized.

[0119] like (The main peak is slightly to the left) and If the area of ​​interest is determined to be a large area of ​​normal skin, the process is terminated. The critical value for determining the burn / normal area, representing the normalized position of the main peak.

[0120] S204: Mask generation and post-processing.

[0121] S204-1: Connected component and area filtering (remove small independent fragments and retain large regions with small effective masks).

[0122] First, calculate the connected components to obtain the area of ​​each component. Maximum connected area Define the adaptive area threshold:

[0123] (16);

[0124] in, Minimum number of pixels to retain; This is the relative threshold coefficient for the area of ​​the connected components, which is related to the maximum object size.

[0125] like If it is not a valid mask, then delete it.

[0126] S204-2: Effective Mask Identification and Conditional Filling

[0127] Identify background connected components, if a certain background component If the mask is not connected to the image boundary, it is considered a "valid mask" with an area of ​​[area value missing]. .

[0128] Fill conditions:

[0129] (17);

[0130] in, This represents the preset absolute lower limit of the "smallest effective mask area". Represents the relative threshold coefficient of the mask hole area; Representing the The area (in pixels) of each background connected component (i.e., candidate hole); This represents the area threshold used to determine whether to fill the void.

[0131] S204-3: Morphological closing operation (filling gaps, merging nearest neighbor fragments).

[0132] After effective mask recognition and filling are completed, morphological closing operations are performed to further process the slits and fragmented regions.

[0133] S205: False color image of the burn area.

[0134] Each pixel within the masked area is rendered in pseudo-color based on peak time, and this is used as the final analysis result.

[0135] In summary, this invention achieves burn region segmentation based on contrast agent blood perfusion burn data acquisition. Compared with traditional burn region segmentation that relies on image or video data streams, it does not depend on the texture, structure, color, or other features of the burned skin, does not require a large dataset or feature classification, has lower computational costs, and does not require the acquisition of additional data information based on the contrast agent blood perfusion burn data acquisition system. It can make full use of the fluorescence intensity and time information of blood perfusion to more accurately distinguish burn and non-burn areas.

[0136] Figure 5 A skin burn analysis system based on contrast agent blood perfusion is shown, comprising:

[0137] The data acquisition unit 501 is configured to acquire the fluorescence sequence of the skin region to be analyzed, the fluorescence sequence containing fluorescence intensity data over time acquired under excitation light source irradiation after injection of contrast agent;

[0138] Preprocessing unit 502 is configured to: perform region of interest cropping and downsampling on the fluorescence sequence to generate a preprocessed fluorescence sequence;

[0139] The curve analysis unit 503 is configured to: extract the fluorescence intensity time curve of each pixel based on the preprocessed fluorescence sequence, and determine the fluorescence intensity peak time corresponding to each pixel;

[0140] The threshold selection unit 504 is configured to: count the peak fluorescence intensity time of all pixels in the preprocessed fluorescence sequence, construct a peak fluorescence intensity time histogram, and select a threshold to distinguish between normal skin and burned skin based on the peak fluorescence intensity time histogram;

[0141] The initial screening mask unit 505 is configured to: perform binarization classification of pixels in the preprocessed fluorescence sequence using a threshold to generate an initial burn area mask;

[0142] Post-processing unit 506 is configured to perform connected component analysis and morphological post-processing on the initial burn area mask to generate a target burn area mask.

[0143] Visualization unit 507 is configured to generate a pseudo-color image of the burn area based on the target burn area mask and the peak time of fluorescence intensity.

[0144] It is understood that the aforementioned units can be individually or entirely merged into one or more other units, or some of the units can be further divided into multiple functionally smaller units. This achieves the same operation without affecting the technical effects of the embodiments of the present invention. The aforementioned units are based on logical functional division. In practical applications, the function of one unit can be implemented by multiple units, or the function of multiple units can be implemented by one unit. In other embodiments of the present invention, the system may also include other units. In practical applications, these functions can also be implemented with the assistance of other units, and can be implemented collaboratively by multiple units.

[0145] According to another embodiment of the present invention, the system of this embodiment can be constructed by running a computer program (including program code) capable of performing the steps involved in the corresponding method of the present invention on a general-purpose computing device, such as a computer, which includes processing elements and storage elements such as a central processing unit (CPU), random access memory (RAM), and read-only memory (ROM). The computer program can be recorded on, for example, a computer-readable recording medium, loaded into the aforementioned computing device through the computer-readable recording medium, and run therein.

[0146] Figure 6A computer device is shown, which includes a processor 601, a communication interface 602, and a computer-readable storage medium 603. The processor 601, communication interface 602, and computer-readable storage medium 603 can be connected via a bus or other means.

[0147] The communication interface 602 is used to receive and send data. The computer-readable storage medium 603 can be stored in the memory of the electronic device. The computer-readable storage medium 603 is used to store computer programs, which include program instructions. The processor 601 is used to execute the program instructions stored in the computer-readable storage medium 603.

[0148] The processor 601 is the computing and control core of an electronic device. It is suitable for implementing one or more instructions, specifically for loading and executing one or more instructions to achieve the corresponding method flow or corresponding function.

[0149] Processor 601 is configured to perform the following procedure:

[0150] Obtain the fluorescence sequence of the skin area to be analyzed. The fluorescence sequence contains fluorescence intensity data over time collected under excitation light source irradiation after injection of contrast agent.

[0151] The region of interest of the fluorescence sequence is cropped and downsampled to generate a preprocessed fluorescence sequence;

[0152] Based on the preprocessed fluorescence sequence, the fluorescence intensity-time curve of each pixel is extracted, and the peak fluorescence intensity time of each pixel is determined.

[0153] The peak fluorescence times of all pixels in the preprocessed fluorescence sequence were statistically analyzed, a peak fluorescence time histogram was constructed, and a threshold for distinguishing normal skin from burned skin was selected based on the peak fluorescence time histogram.

[0154] The pixels in the preprocessed fluorescence sequence are binarized and classified using a threshold to generate an initial burn area mask;

[0155] Connectivity analysis and morphological post-processing are performed on the initial burn area mask to generate the target burn area mask.

[0156] A pseudo-color image of the burn area is generated based on the mask of the target burn area and the peak time of fluorescence intensity.

[0157] This invention also provides a computer-readable storage medium, which is a memory device in an electronic device for storing programs and data. It is understood that the computer-readable storage medium here may include both built-in storage media in the electronic device and extended storage media supported by the electronic device. The computer-readable storage medium provides storage space for storing the processing system of the electronic device.

[0158] Furthermore, this storage space also contains one or more instructions suitable for loading and execution by the processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory; alternatively, it can also be at least one computer-readable storage medium located remotely from the aforementioned processor.

[0159] In one embodiment, the computer-readable storage medium stores one or more instructions; the processor loads and executes the one or more instructions stored in the computer-readable storage medium to perform the following process:

[0160] Obtain the fluorescence sequence of the skin area to be analyzed. The fluorescence sequence contains fluorescence intensity data over time collected under excitation light source irradiation after injection of contrast agent.

[0161] The region of interest of the fluorescence sequence is cropped and downsampled to generate a preprocessed fluorescence sequence;

[0162] Based on the preprocessed fluorescence sequence, the fluorescence intensity-time curve of each pixel is extracted, and the peak fluorescence intensity time of each pixel is determined.

[0163] The peak fluorescence times of all pixels in the preprocessed fluorescence sequence were statistically analyzed, a peak fluorescence time histogram was constructed, and a threshold for distinguishing normal skin from burned skin was selected based on the peak fluorescence time histogram.

[0164] The pixels in the preprocessed fluorescence sequence are binarized and classified using a threshold to generate an initial burn area mask;

[0165] Connectivity analysis and morphological post-processing are performed on the initial burn area mask to generate the target burn area mask.

[0166] A pseudo-color image of the burn area is generated based on the mask of the target burn area and the peak time of fluorescence intensity.

[0167] The present invention also provides a computer program product or computer program comprising computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the following process:

[0168] Obtain the fluorescence sequence of the skin area to be analyzed. The fluorescence sequence contains fluorescence intensity data over time collected under excitation light source irradiation after injection of contrast agent.

[0169] The region of interest of the fluorescence sequence is cropped and downsampled to generate a preprocessed fluorescence sequence;

[0170] Based on the preprocessed fluorescence sequence, the fluorescence intensity-time curve of each pixel is extracted, and the peak fluorescence intensity time of each pixel is determined.

[0171] The peak fluorescence times of all pixels in the preprocessed fluorescence sequence were statistically analyzed, a peak fluorescence time histogram was constructed, and a threshold for distinguishing normal skin from burned skin was selected based on the peak fluorescence time histogram.

[0172] The pixels in the preprocessed fluorescence sequence are binarized and classified using a threshold to generate an initial burn area mask;

[0173] Connectivity analysis and morphological post-processing are performed on the initial burn area mask to generate the target burn area mask.

[0174] A pseudo-color image of the burn area is generated based on the mask of the target burn area and the peak time of fluorescence intensity.

[0175] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can implement the described functions using different methods for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0176] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic cable, digital cable) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data processing device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0177] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for analyzing skin burns based on contrast agent blood perfusion, characterized in that, Includes the following steps: Obtain the fluorescence sequence of the skin region to be analyzed, wherein the fluorescence sequence includes fluorescence intensity data over time collected under excitation light source irradiation after injection of contrast agent; The fluorescence sequence is cropped and downsampled to generate a preprocessed fluorescence sequence; Based on the preprocessed fluorescence sequence, the fluorescence intensity-time curve of each pixel is extracted, and the peak fluorescence intensity time corresponding to each pixel is determined; The peak fluorescence times of all pixels in the preprocessed fluorescence sequence are statistically analyzed, a peak fluorescence time histogram is constructed, and a threshold for distinguishing normal skin from burned skin is selected based on the peak fluorescence time histogram. The threshold is used to perform binarization classification on the pixels in the preprocessed fluorescence sequence to generate an initial burn area mask. Connectivity analysis and morphological post-processing are performed on the initial burn area mask to generate the target burn area mask; Based on the target burn area mask and the fluorescence intensity peak time, a pseudo-color image of the burn area is generated; The steps in constructing the fluorescence intensity peak time histogram also include: The fluorescence intensity peak time histogram is smoothed and denoised to generate a smoothed histogram curve; Local extremum detection is performed based on the smoothed histogram curve to identify the set of peaks and valleys. Calculate the significance coefficient of each peak in the peak set, remove weak peaks with significance coefficients lower than a preset minimum significance coefficient, and obtain the main peak set; In the step of selecting the threshold for distinguishing between normal skin and burned skin, if the set of main peaks contains two main peaks, the following operations are performed: calculate the relative valley depth between the two main peaks, determine whether the relative valley depth meets the preset acceptance threshold condition, and if the acceptance threshold condition is met, determine the time value corresponding to the valley point between the two main peaks as the threshold. If the set of main peaks contains only one main peak or does not meet the double-peak determination condition, then perform the following operations: determine the nearest valley points to the left and right of the main peak position as candidate valleys, compare the depth of the candidate valleys relative to the main peak, select the time value corresponding to the candidate valley with the more significant depth as the threshold, if the region of interest is determined to be a large area of ​​burns, then directly retain the entire region as the initial burn area mask, if the region of interest is determined to be a large area of ​​normal skin, then terminate the process.

2. The skin burn analysis method based on contrast agent blood perfusion as described in claim 1, characterized in that, In the step of generating the target burn area mask, the connected component analysis specifically includes: Calculate the area of ​​each connected component in the initial burn area mask, and determine the maximum connected area; Based on the maximum connected area and the preset minimum number of retained pixels, calculate the adaptive area threshold; Delete connected components whose area is smaller than the adaptive area threshold and which are determined to be invalid masks.

3. The skin burn analysis method based on contrast agent blood perfusion as described in claim 1, characterized in that, In the step of generating the target burn area mask, the morphological post-processing specifically includes: Identify background connected components; if a background connected component is not connected to the image boundary, it is determined to be a valid mask hole. When the area of ​​the effective mask hole is smaller than the preset small effective mask area, the effective mask hole is filled. A morphological closing operation is performed on the filled mask to eliminate regional gaps and merge neighboring fragments, thus obtaining the target burn area mask.

4. The method for analyzing skin burns based on contrast agent blood perfusion as described in claim 1, characterized in that, The steps for generating a pseudo-color image of the burn area specifically include: Traverse every pixel within the mask range of the target burn area; Read the peak fluorescence intensity time corresponding to each pixel in the fluorescence intensity-time curve; The peak fluorescence intensity time is mapped to a specific pseudo-color value, and the pixels within the mask range of the target burn area are rendered and colored.

5. A skin burn analysis system based on contrast agent blood perfusion, characterized in that, include: The data acquisition unit is configured to acquire a fluorescence sequence of the skin region to be analyzed, the fluorescence sequence including fluorescence intensity data over time acquired under excitation light source irradiation after injection of contrast agent; The preprocessing unit is configured to: perform region-of-interest (ROI) cropping and downsampling on the fluorescence sequence to generate a preprocessed fluorescence sequence; The curve analysis unit is configured to: extract the fluorescence intensity time curve of each pixel based on the preprocessed fluorescence sequence, and determine the fluorescence intensity peak time corresponding to each pixel; The threshold selection unit is configured to: count the peak fluorescence times of all pixels in the preprocessed fluorescence sequence, construct a peak fluorescence time histogram, and select a threshold to distinguish between normal skin and burned skin based on the peak fluorescence time histogram; The initial screening mask unit is configured to: perform binarization classification of pixels in the preprocessed fluorescence sequence using the threshold to generate an initial burn area mask; The post-processing unit is configured to perform connected component analysis and morphological post-processing on the initial burn area mask to generate a target burn area mask. The visualization unit is configured to generate a pseudo-color image of the burn area based on the target burn area mask and the fluorescence intensity peak time. The steps in constructing the fluorescence intensity peak time histogram also include: The fluorescence intensity peak time histogram is smoothed and denoised to generate a smoothed histogram curve; Local extremum detection is performed based on the smoothed histogram curve to identify the set of peaks and valleys. Calculate the significance coefficient of each peak in the peak set, remove weak peaks with significance coefficients lower than a preset minimum significance coefficient, and obtain the main peak set; In the step of selecting the threshold for distinguishing between normal skin and burned skin, if the set of main peaks contains two main peaks, the following operations are performed: calculate the relative valley depth between the two main peaks, determine whether the relative valley depth meets the preset acceptance threshold condition, and if the acceptance threshold condition is met, determine the time value corresponding to the valley point between the two main peaks as the threshold. If the set of main peaks contains only one main peak or does not meet the double-peak determination condition, then perform the following operations: determine the nearest valley points to the left and right of the main peak position as candidate valleys, compare the depth of the candidate valleys relative to the main peak, select the time value corresponding to the candidate valley with the more significant depth as the threshold, if the region of interest is determined to be a large area of ​​burns, then directly retain the entire region as the initial burn area mask, if the region of interest is determined to be a large area of ​​normal skin, then terminate the process.

6. A computer device, characterized in that, include: Processor and computer-readable storage media; A processor, adapted to execute computer programs; A computer-readable storage medium storing a computer program that, when executed by the processor, implements the skin burn analysis method based on contrast agent blood perfusion as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted to be loaded by a processor and executed as described in any one of claims 1 to 4 for the analysis of skin burns based on contrast agent blood perfusion.

8. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the skin burn analysis method based on contrast agent blood perfusion as described in any one of claims 1 to 4.

Images