Procedure for assessing areas of interest on a skin surface

Abstract

A method for assessing multiple areas of interest on the skin surface of an object. The method comprises obtaining one or more first images of at least a part of the skin surface; identifying multiple areas of interest in the one or more first images; generating multiple second images based on the obtained one or more first images, where each of the multiple second images comprises one of the multiple identified areas of interest, respectively; assigning at least one feature to each of the multiple identified areas of interest in each of the multiple second images, respectively; classifying the multiple second images based on at least one feature; obtaining one or more third images of at least one of the multiple identified areas of interest based on the classification of the multiple second images;associating one or more third images with the multiple second images; assessing at least one of the multiple identified areas of interest in the multiple second images and / or the one or more third images based on the classification.;

Description

The VECTRA® WB360 offers an innovative solution with 92 high-resolution cameras that use polarized light to image the entire body in 0.3 seconds under standardized conditions. This system creates a detailed 3D image of the patient's skin surface using stereophotogrammetry. The accompanying Vectra DermaGraphix software automatically identifies and characterizes visible skin lesions larger than 5 2 mm. The use of the VECTRA® WB360 in combination with 3D-CNN can reduce the number of skin lesions that require further examination by 75%. This leads to significant time savings and increased efficiency. However, due to the lower specificity of the 3D-TBPCNN, there is a risk of overtreatment and it can cause anxiety in the patient. For this reason, there is a need for improved techniques and / or methods that can increase specificity without compromising sensitivity. to reduce in order to improve the accuracy of the diagnosis, thereby reducing unnecessary treatments and increasing the patient's peace of mind.US2016 / 275681A1 describes a method and setup that helps a user, such as a physician, to quickly and effectively examine large skin surfaces by determining an attribute associated with each of the multitude of skin features included in one or more images of the skin; generating a tile image of each of the multitude of skin features; arranging the tile images in accordance with the attribute associated with each of the multitude of skin features; and controlling a display device to display the tile images of the multitude of skin features. 20 Summary of the invention Forms of the invention aim to provide a method for assessing multiple areas of interest on the surface of an object. The method involves obtaining one or more first images of at least part of the surface, identifying multiple areas of interest in the one or more first images, generating multiple second images based on the obtained one ormultiple first images, where each of the multiple second images respectively comprises one of the multiple identified areas of interest, assigning at least one attribute to each of the multiple identified areas of interest in respectively each of the multiple second images, classifying the multiple second images based on 30 the at least one attribute, obtaining one or more third images of at least one of the multiple identified areas of interest based on the classification of multiple second images, associating the one or more third images with multiple second images, evaluating at least one of the multiple identified areas of interest in multiple second images and / or the one or more third images based on the classification. BE2025 / 5019 Preferably, the method comprises determining the status of at least one of the multiple areas of interest based on the assessment of multiple areas of interest.Preferably, the acquisition of multiple second images based on the obtained one or more first images and the cropping of the one or more first images. Preferably, it includes at least one feature of a size, shape, color, location, symmetry, medical relevance, pathological classification of multiple identified areas of interest. Preferably, the classification includes at least one of the sorting, grouping, or filtering of the multiple second images. Preferably, the classification of multiple second images and the classification of the multiple second images into two or more hierarchical levels, where each successive level includes two or more subgroups of each group multiple second images of a preceding level. Preferably, the multiple identified areas of interest are classified in such a way that it facilitates the assessment of the identified areas of interest. Preferably, the method involves assigning a color code to at least one of severalareas of interest based on the assessment of several identified areas of interest. 25 Preferably, the color code corresponds to the state of at least one of the several areas of interest. Preferably, the classification and / or assessment of at least one of the identified areas of interest is performed by a computer program. 30 Preferably, the method includes displaying the classification of several second images BE2025 / 5019 Preferably, one or more of the images includes at least one dermatoscopic image, confocal microscopy image, or optical coherence tomography (OCT) image. Preferably, the method includes generating a three-dimensional, 3D, model of at least a part of the skin surface based on the obtained one or more first images. Preferably, the method includes displaying each of the several second images on the 3D model. 10 Preferably includes displaying each of the multiple second images on the 3D modellinking each of the multiple or second images to a location on the 3D model. Preferably, the method involves processing at least one of the one or more first images, where the execution includes at least one of a filter procedure, a noise removal procedure, or a hair removal procedure. Preferably, the method involves performing a correction of multiple or second images, comprising at least one of a color correction with respect to a general skin tone or a perspective correction. Preferably, the multiple areas of interest are identified by means of an inspection. Preferably, the method is executed at at least one first time point and a second time point such that at least one feature includes a change or a rate of change in the at least one feature. Short figure description The above and other advantageous properties and purposes of the invention will become clearer and the invention will be better understood based on the following detailed 30description when read in combination with the drawings in the appendix, in which: FIG.1 shows a schematic representation of an example system in accordance with the present invention; FIG.2 shows a flowchart that represents an example method in accordance with the present invention;35 BE2025 / 5019 FIG.3 shows an example representation of a visualisation scheme that is implemented with examples of the present invention; FIG.4 shows a further example representation of a visualisation scheme that is executed with an example execution of the present invention; FIG.5 shows a further example representation of a visualisation scheme that is executed with an example execution of the present invention; and FIG.6 shows a schematic representation of an example method in accordance with the present invention. Detailed forms of execution10 The following detailed description focuses on certain specific forms of execution, however, the doctrine therein is applied in different ways.The present invention will be described with respect to specific embodiment and the invention is however not limited thereto, but only by the conclusions. As used herein, comprising the singular form "a", "the" and "the", although the singular and plural references on the side context clearly dictates otherwise. The terms "comprehensive", "includes" and "composed of" as used herein are synonymous with "inclusive", "includes" or "containing", "contains". The terms "comprehensive", "includes" and "composed of", when referring to said components, elements or procedure steps, also include execution forms that "consist of" the components, elements or procedure steps.20 Furthermore, the terms first, second, third and further descriptions in the conclusions are used to distinguish between comparable elements and not necessarily for describing a consecutive or chronological order, unless this is specified. It is clear that the terms thus used are interchangeable undersuitable circumstances and that the forms of execution of the invention described herein can work in a different order than described or illustrated herein. Reference in this specification to "one version", "one version", "some aspects", "one aspect" or "one aspect" means that a specific feature, structure or feature described in connection with the version or aspect is included in at least one form of execution of the present invention. The forms of appearance of the phrases "in one version", "in one version", "some aspects", "one aspect" or "one aspect" at different places in this specification therefore do not necessarily all refer to the same version or aspects. Furthermore, the specific features, structures or features may be combined in any suitable manner, as will be clear to a skilled professional in this field, in one or more forms or aspects. Furthermore, although some forms or aspects described herein include some but no other features in other forms or BE2025 / 5019aspects are included, combinations of characteristics of different forms of execution or aspects intended to fall within the context of the invention and to form different forms of execution or aspects, as would be understood by the skilled person. In the attached claims, for example, all characteristics of the claimed forms of execution or aspects can be used in any combination. 5 Thus, for example, skilled people will understand that the diagrams contained herein are conceptual representations of illustrative structures that embody the principles of the invention. In addition, skilled people will understand that flowcharts represent such different processes that can be broadly represented on a computer-readable medium and can be executed by a computer or processor, 10 regardless of whether such a computer or processor is explicitly shown. The functions of the various elements shown in the images, including all functional blocks that are designated as “processor” or “processing”, can beexecuted using specific hardware and hardware capable of executing software in combination with suitable software. When functions are executed by a processor, they may be executed by a single specific processor, by a single shared processor, or by a large number of separate processors, some of which are shared. Furthermore, the explicit use of the term “processor” or “controller” should not be interpreted as a reference to exclusively hardware capable of executing software and implicitly may also include, without limitation, hardware for digital signal processors (DSPs), network processors, application-specific integrated circuits (ASICs), user-programmable gate arrays (FPGAs), software storage ROMs, RAMs (random access memory), and non-volatile storage media. Other hardware, conventional and / or custom-made, may also be included. Software modules, or simply modules intended as software, can be in here25are displayed as a combination of flowchart elements or other elements that indicate the execution of process steps and / or textual description. Such modules can be executed by hardware that is explicitly or implicitly displayed. In the drawings, the same reference number is assigned to the same or analogous element. 30 Figure 1 shows an example implementation form of a system100 for obtaining and processing images according to a first aspect of the current description. As shown in Figure 1, the system100 comprises an image registration system110 coupled to a processing module140. The image registration system110 can comprise 2D or 3D image registration elements for obtaining one or more images of an object130 or a part thereof.35 An advantage is that the obtained images can be single or multimodal, including BE2025 / 5019 for example images of standard white light, polarized light and / or fluorescent light. These can be obtained at selected wavelengths and / or illuminated with selected wavelengthsof light. The images obtained by the image registration system 110 are delivered to the processing module 140 during processing, as described in more detail below. An additional advantage is that the processing module 140 can also control the image registration system 110, for example by regulating specific aspects of the image registrations and / or the lighting of the object 130. As shown in Figure 1, the image registration system 110 can comprise one or more light sources which, when activated, illuminate the object 130. This illumination can advantageously comprise one or 10 more respective filter elements. Light reflected by the object 130, which is positioned in the correct manner, can be captured by the image registration system 110 comprising one or more filter elements (not specifically shown). These can comprise one or more filters for transmitting or blocking light of a selected wavelength or band of wavelengths, and / or polarizers (hereinafter collectively referred to as "filters"), which selectively select or exclude 15can be placed on a respective optical path of the filter element. As implied earlier, we note that the term "light" as used here is not necessarily limited to electromagnetic radiation visible to humans. It is noteworthy that "light" as used here refers to electromagnetic radiation in any part of the electromagnetic spectrum, including that above and / or below and / or within the range of human vision. 20 The processing module 140 can be implemented with one or more computers, workstations or the like, which function according to one or more processing programs 145 that are included in a compatible machine-readable medium. As is easy to understand for those skilled in the art, the processing module 140 can be connected to storage module(s) 150 and display module(s) 160 as the implementation requirements prescribe. The 25 processing module140canalso be connected to a communication network170, such as the internet, for sending images and / or data, and / or receiving commands,software updates or the like. The system100 can also include or cooperate with additional instruments, such as a dermatoscope180, which can be used to take (close-up) images of 30 areas of interest on the skin surface of the object130. Such (close-up) images obtained via the dermatoscope can subsequently be advantageously associated with corresponding images obtained by the image registration system110 and / or with the areas of interest identified therein. In addition to (close-up) images, the dermatoscope180 can also include data linked to the 35 images, such as, for example, gyroscope or position data from the dermatoscope180. BE2025 / 5019 These data can help determine the orientation of the obtained dermatoscopy images relative to the object130. It is also possible, especially with a device with processing capabilities, that the dermatoscope180 performs some pre-processing of the images before they are delivered to the processing module140. As mentioned earlier, other instruments can cooperate with the5processing module 140 to provide both images and other data. This includes systems for obtaining (close-up) images, such as confocal microscopy systems, optical coherence tomography (OCT) systems and colorimeters, among others. In an illustrative version, the image registration system 110 is one of the VECTRA® imaging systems from Canfield Scientific, Inc. (for example, a VECTRA® WB180, WB360 or H110 imaging system) and the dermatoscope 180 is a VEOS dermatoscope, also from Canfield Scientific, Inc. The current invention system 100 is specifically designed to obtain and analyze images of the skin surface of object 130. The primary purpose of this is to screen object 130 for possible melanomas. This screening process is suitable for patients with a low to high risk of melanomas. Through this application, the system offers a valuable tool in the early detection and monitoring of skin cancer, which is essential for effective treatment care. It is important to note that the example system 100 is only one ofThe various possible configurations are illustrated in this current description. For example, the different modules of system 100 do not need to be in the same location. Thus, the image registration system 110 and the display module 160 can be located in a dermatologist's practice, while the processing module 140 and the storage module 150 can be at an external location or "cloud-based", and communicate with the image registration system 110 and the display module 160 via the communication network 170. In other example configurations, the display module 160 can be located remotely from the image registration system 110, allowing a dermatologist at the display module 160 to examine a patient's skin remotely. Figure 2 shows a flowchart of an example method200 for processing images according to aspects of this current description. As will be understood by an expert, the example method200 can be carried out using system 10030 shown in Figure 1, where the processing module 140 operates according to the processing programs 145.As shown in Figure 2, and with simultaneous reference to system 100 from Figure 1, the example procedure 200 begins at step 210, in which one or more images of at least a part of the skin surface of the object 130 are obtained by means of the image registration system 110. The one or more images obtained are 2D images. BE2025 / 5019 In addition to merging these images to form a 3D model of the object at step 215, the images are processed and analyzed to identify areas of interest. The identified areas of interest are displayed on the 3D model of the object adapted to its shape. Alternatively, in a 2D registration system implementation, the identified areas of interest are projected onto the 2D images of a body map. In an illustrative execution, one or more 2D images are obtained synchronously to construct a 3D model of the photographed object. This allows every pixel in each source image to be mapped to real-world x, y, z coordinates in 3D space. This is primarilyimportant for handling overlapping 2D images and for accurate dimensional measurements of the object, and thus for every area of ​​interest. The obtained one or more 2D images are used to identify areas of interest, such as skin features. This identification of areas of interest can be advantageously performed by image processing algorithms that automatically segment the boundaries of the identified areas of interest, or by varying degrees of user involvement, whether or not under the supervision of a practitioner. In the example method, the identification of multiple areas of interest in the one or more initial images is performed in step 230, which preferably includes a preprocessing or filtering step 220 or is preferably preceded by a preprocessing or filtering step to remove noise or other distracting elements in the image, such as hair, that could otherwise negatively affect segmentation. In example implementations, filtering 220 and / or the identification of areas of interest can230selectively be executed in response to user input235. For example, the threshold values ​​used to identify areas of interest230 can be adjusted by user input to more accurately regulate the relative sensitivity and specificity of the identification of areas of interest. As is easy to understand, some25 users may choose to increase sensitivity to increase the number of true positives, while others may wish to decrease sensitivity to reduce the number of false positives. In this way, a user who desires non-standard functionality can adjust the threshold as desired. Additionally, in example setups, a user can also manually remove identified areas of interest if some have been incorrectly identified, for example by noise,30 or even redraw the boundary of the area of ​​interest. If the boundary of an area of ​​interest is redrawn, all measurements or metric values ​​affected by this arepreferably regenerated based on the revised boundary, whereby all displays with that area of ​​interest are also updated. Areas of interest identified in overlapping areas between or across multiple 35 first images are merged in entry 240 to ensure that individual BE2025 / 5019 areas of interest obtained in multiple images are not reported twice. This can be done using a suitable image correspondence method, involving the 3D model if available. In an example implementation, generating the 3D model in step 215 involves generating a series of non-overlapping masks corresponding to a series of obtained 2D images. After segmentation has been performed on each 2D image in its entirety in step 230, the corresponding masks are applied in step 240 to eliminate any overlap. Alternatively, images from a camera placed at the best angle to view a specific area of ​​the skin can be selected for the segmentation ofareas of interest within that area. In this case, each area is analyzed once, specifically within the image that offers the best view of that area. Overlapping areas in other images that are not as optimal are ignored. As another alternative, areas of interest can be detected in all 2D images, regardless of the view, and subsequently reprojected into other images based on pixel-to-3D coordinate adjustments, whereby the union, intersection, or another suitable combination of the segmented areas of interest is retained. In Step 250, the identified areas of interest can subsequently be isolated, or cut from the original 2D images, to create a second image or thumbnail view of each area of ​​interest. Perspective correction can also be applied to these second images, if available, so that it appears as if the second images have been captured perpendicular to the area of ​​interest. The second images can be cropped into a square, circle, rectangle, hexagon, or any other suitable shape.In step 255, a color correction can be applied to the second images of the areas of interest to normalize the background skin of the object around the areas of interest. When performing such a color correction, an estimate of the general skin color of the object can first be calculated based on the original captured images. Subsequently, the area of ​​the background skin within each cropped second image (outside the identified area of ​​interest) can be used to calculate a color correction function for that entire second image with respect to the skin color estimate of the object. For example, each second image can be shifted in the L*a*b* color space such that the delta between the skin color of the person and the background skin color of the second image averages zero. The L*a*b color space expresses a color as three values: L* stands for brightness, a* stands for the red / green axis, and b* stands for the yellow / blue axis. Each second image can also be corrected based on the distance or position of the object in the second image.images in relation to the deposition of the light sources. Thus, elements located closer to a flash can be made darker, while elements further away from a flash can be made lighter. An advantage of color correction of the individual second images with respect to the skin color of the lens is a more consistent representation. A viewer can be distracted by intensity shifts in cropped second images as a result of factors such as distance to ambient light sources, flashes, or even tanning. By normalizing the background color of the skin for each second image, it becomes easier to compare the areas of interest at different locations with each map. Instap260, for each identified area of ​​interest, a series of one or more characteristics can be assigned that describe that identified area of ​​interest, such as, for example, size (e.g., area, diameter, main / mid-axis length, circumference), intensity, color (e.g., average RGB or L*a*b* values, average delta RGB or L*a*b* values ​​relative to 10of the local skin color background), symmetry, irregularity of the edge, circularity, eccentricity, location (e.g., x-, y-, z-coordinates in real-world space, anatomical body part such as head, neck, fore or hind body, left or right arm, etc.), a derived measure, such as “pathological significance” or a pathological classification, among other possibilities. For each attribute relating to a physical measurement, such as area, length, width, etc.,15 the actual pixel-based measurement of the 2D image is preferably converted into a physical measurement (such as in inches, or millimeters, etc.) using spatial information, such as from a ruler or other reference object of known size that appears in the image. In the case of 3D photography, similar information can be derived from a calibrated 3D model reconstruction. The advantage is that with the information from the 3D model, areas of interest can be corrected for perspective distortion of the 2D projection, which is noticeably curved.surfaces—before their characteristics are calculated. This enables very accurate measurements, despite the fact that 2D images are not captured perpendicular to the surface of the object or where the number of pixels per inch (PPI) can vary even within the same image because different areas of the object are located at different positions and distances from the camera. In this step, the second images generated are classified into a visualization or display scheme to obtain a simple overview at once of the visual aspects of the areas of interest represented by the second images. Illustrative representations of such a visualization scheme are shown in Fig. 3 and Fig. 4.30 Second images can be square, circular, hexagonal, or another suitable shape and can be grouped together in a grid, row, column, spiral, tree, or other suitable layout, as shown respectively in Figs. 3 and 4. In addition, the second images can be sorted or arranged according to one or more characteristics or combinations.of that. For example, in the illustrative setup shown in Fig. 3, the circular second images of identified areas of interest are classified in a spiral sorted on the basis of area. The second images can also be grouped on the basis of one or more characteristics of the areas of interest shown in the second images or divided into groups on the basis of a set of criteria or threshold values ​​for one or more characteristics. These groups can then be arranged together on a screen, so that a viewer can see which areas of interest have similar characteristics and which may be outliers. In the example setup in Fig. 4, the second images of the same identified areas of interest as in Fig. 3 are arranged in groups on the basis of area sizes. These groups can also be arranged in a hierarchy of an arbitrary number of levels, where each lower level divides the upper group on the basis of one or more characteristics of theareas of interest. Thus, each of the groups in Fig. 4 can be further subdivided into subgroups based on, for example, color, regularity of the edges, location of the body, or any other suitable characteristic. Areas of interest that cover the entire body can be grouped, or they can be organized by grouping parts of the body, such as the face, the trunk, legs, arms, etc. A system according to the current description preferably enables a user to control one or more settings that influence the presentation of areas of interest, as illustrated in FIGS.3 and 4. Preferably, the user can select one or more filters20 to apply to the characteristics of the areas of interest to limit the areas of interest to a subgroup in which the user is particularly interested. For example, the user can select only the areas of interest on the left arm with an area of ​​more than 5 mm². In response, the system updates the displayed visualization so that only the areas of interest that meet these criteria are shown or highlighted.25The location of the identified areas of interest (such as the position of the center or the boundary, etc.) can be projected back onto the original 2D images or the 3D model in step 275, for an interactive view of the object with the identified areas of interest. This can happen before or after the selective filtering described above. Such visualizations are illustrated in FIGS.3 and 4, respectively at 320 and 420.30 The advantage is that areas of interest projected back onto the original 2D images or the 3D model can optionally be color-coded based on one or more characteristics of each area of ​​interest. In this way, a user can visualize the actual position of the areas of interest and how the areas of interest actually appear on the subject, together with additional information and notes to help him make sense of what he sees and identify the most interesting areas of interest. These BE2025 / 5019 areas of interest can also be marked in step280 to be recorded in a thirdimage with an alternative image capture system, such as dermoscope, confocal microscopy, optical coherence tomography (OCT), or another close-up image capture system. Once these three images have been captured, they can be linked to their respective areas of interest, so that the user can easily mark an area of ​​interest and view the additional information and images in a consolidated manner. Such three images and information are shown in FIGS. 3 and 4, respectively at 320 and 420. As can be seen in FIGS.3 and 4, information regarding individual areas of interest can be displayed in 320 and 420. FIG.3 shows the example case of an area of ​​interest F1, which is marked in display area 310 and for which additional information10 is displayed in display area 320. Area of ​​interest F1 can be selected, for example, by a user who clicks on the second image displayed in area 310 of area of ​​interest F1 or selects it in another way.In the example setup in Fig. 3, display area 320 comprises parts 321-323. Display part 321 shows an image of the body part in which area of ​​interest F1 is found, 15 in this case the torso, with area of ​​interest F1 marked, for example by a circle or another suitable means by which a user can easily find area of ​​interest F1 in the image. Display part 322 shows one or more second and / or third images of the area of ​​interest F1 itself. As shown in Fig. 3, display part 322 contains a conventional reflection image of the area of ​​interest F1 and a version of the image with additional graphic information on it, in this case a contour map in which the contour lines represent different intensity levels (L*) within the area of ​​interest. Additional images of the area of ​​interest F1 that were created in different modalities (e.g. polarized, UV) and / or at different times can also be displayed. The image(s) displayed in part 322 can be obtained from the original image of the body part or the whole body orfrom other sources, such as dermatoscope 180 in FIG.1. Part 323 of display area 320 contains alphanumeric information regarding area of ​​interest F1, such as an ID, location, size metric (e.g., area, length of the main axis and the minor axis, perimeter), shape metric (e.g., circularity, eccentricity), and color information (e.g., average L*a*b* values ​​within the area of ​​interest, delta E*). Delta E* can be calculated by: ΔE* = √{square root of ((L*FL*S)² + (a*Fa*S)² + (b*Fb*S)²)} where L*F, a*F and b*F represent the average L*a*b* values ​​of the area of ​​interest and L*S, a*S and b*S represent the average L*a*b* values ​​of the local background skin within the second image but outside the area of ​​interest. 35 BE2025 / 5019 In addition to the views as shown in 310 and 410 in FIGS. 3 and 4, an example system can display a whole or partial body view according to the current description, as shown in FIG. 5 at 510. In such a view, the system canoffer selection tool515 that a user can move to a displayed area of ​​interest to view information about the area of ​​interest. An enlarged view of the area of ​​interest can be viewed within the selection tool, which simulates a magnifying glass bicycle or similar, and / or within an area520. Alphanumeric information about the area of ​​interest can be displayed near the selection tools and / or in area520. The views in Fig. 5 are also useful for implementations where the system skips generating, classifying, and displaying classified second images (as shown in Figs. 3 and 4) and instead displays one or more selected areas of interest directly on the whole or partial body view in 510 and / or 520. The areas of interest can be selected based on one or more features, such as, for example, the areas of interest with an pathological significance greater than a threshold value. The areas of interest on the skin thatif the selection criteria are met, they can be displayed by the system with an indication, or they can otherwise be marked, for example by a circle or another suitable indicator with which a user can easily locate the area or areas of interest in the image. In step 290, one or more of the areas of interest are assessed based on the second and / or third images. Preferably, the assessment is performed on the basis of the classification. Such an assessment of one or more areas of interest can be performed by a user. Alternatively, the assessment of one or more areas of interest can be performed by an automated procedure using a computer program. Based on the assessment, the selection can be refined. This offers the advantage that the sensitivity and specificity of the selection are increased to provide an accurate selection. Based on the assessment of the selected areas of interest, a state is assigned to the areas of interest. This state can be displayed with acolor code or similar to indicate what type of attention is required for each of the areas of interest. For example, whether the area of ​​interest must be followed up within 6 months or within the year, or if no further follow-up is needed.30 In this way, areas of interest are automatically identified on the skin; one or more of their characteristics are determined (e.g. measured, classified); one or more of the identified areas of interest are selected on the skin based on one or more characteristics; the selected areas of interest, apparently those that deserve further attention, are marked to bring them to the user's attention; the selected areas of interest are further35 assessed to guarantee the accuracy of the selection; and based on the assessment, the BE2025 / 5019 status of the areas of interest is determined and indicated to inform the user about the type of attention that each area of ​​interest requires. FIG. 6 further shows a schematic representation of the flowchart of the example procedure shown in FIG. 2. FIG. 6 shows obtaining one or more firstimages of a part of the skin surface210. In one or more first images obtained, several areas of interest are subsequently identified230. Several second images are generated based on the obtained one or more first images, whereby each of several second images respectively comprises one of several identified areas of interest250. At least one characteristic is assigned to each of several identified areas of interest in respectively each of several second images10 260. Based on the assigned characteristic, several second images can be classified270. Preferably, based on the classification, a third image of at least one of several areas of interest is obtained. Preferably, the third image is a close-up image of at least one area of ​​interest. Alternatively, the one or more third images are subsequently associated with the corresponding second images.15 For areas of interest for which information is available at multiple points in time,For example, if two or more images or sets of images are captured at different points in time, identified areas of interest can be tracked from one point in time to another via various methods, such as image recording or tracking the descriptors of areas of interest. If the same area of ​​interest can be localized at one or more points in time, changes in the characteristics of the area of ​​interest can be used to sort, group, or filter the visualization scheme. For instance, changes in the area of ​​selected areas of interest over the course of a certain period can be determined and used to generate a spiral chart, such as that of 310, in which the areas of interest are ranked based on the percentage increase in area. Such a view enables a user to quickly see which areas of interest have changed the most over a certain period, thereby justifying closer inspection. In additionIn addition to arranging the second images of areas of interest in accordance with one or more of their static characteristics (i.e., characteristics determined at a single point in time), embodiments of the current publication may also base such arrangements on one or more dynamic characteristics (e.g., changes or rates of change of characteristics). For display in a visualization scheme (bubble / spiral chart, histogram view, etc.), the second images of the areas of interest are preferably color-corrected as described above in Fig. 2. In a sample implementation, a map with areas of interest can be provided by the system 100 to 35 the dermatoscope 180 to indicate the location of the areas of interest for which the third BE2025 / 5019 images must be taken with the dermatoscope 180. The map can be displayed on the dermatoscope 180 or screen 160, where areas of interest selected for additional imaging are marked. If third dermatoscopic images aremade with dermatoscope180, the images can easily be linked to the relevant areas of interest and passed on to system100.5 As noted earlier, one characteristic of each area of ​​interest is the anatomical location, namely the body part or part of the body where the characteristic is located. Such information regarding the anatomical location can be determined in various manual or automated ways. For example, for an image of an individual body part or region, features10 detected in the image are classified in accordance with the location on the body that corresponds to that image; for example, the location of the body part of a feature detected in an image of the left arm is determined as the left arm. Such an association between the feature, the body part or the location can be made manually by a user.15 Alternatively, the determination of the anatomical location(s) of features can be performed by an automated procedure. In such a procedure, an anterioror posterior silhouette projection of the body entered into a classification method for body parts. Such a method can be implemented, for example, with a neural network trained to classify the boundary coordinates of the silhouette into the respective body parts, a rule-based set of etheuristics (for example, the legs are always located below the torso for a standing person), or a combination of techniques. Once the main body parts have been identified and segmented, identified areas of interest can be reprojected and compared with a body part map to automatically label each area of ​​interest with its anatomical location. With automated marking techniques, information from the 3D model can also be used, for example, to distinguish the anterior or posterior regions based on x-, y-, etc. coordinates of the area of ​​interest relative to the known coronal plane or based on the direction of thearea normal vector at the location of the area of ​​interest on the 3D model. Alternative marking procedures can operate entirely in 3D, such as through close correspondence matching.30 Based on the description above, the skilled person will understand that the invention can be implemented in different ways and on the basis of different principles. In doing so, the invention is not limited to the forms of implementation described above. The implementation forms described above, as well as