Computer-based method, diagnostic and medical device and GUI for analyzing multiple image

Abstract

A first aspect of the present disclosure relates to an image analysis method comprising the steps of:-determining a first set of requirements; acquiring a plurality of images of the sample from an imaging device, in particular a microscope, according to the first set of requirements; -acquiring a second set of requirements; -analysing the samples in the plurality of images according to the second set of requirements.

Description

Technical Field

[0001] The present disclosure relates to computer-based methods for analyzing a plurality of images, and also discloses diagnostic and / or medical devices that can be used with the methods. In addition, a GUI for use with the embodiments is also introduced. Background Art

[0002] Biological, chemical, diagnostic or medical analysis based on samples or sample image streams uses specialized imaging techniques to examine and interpret the structure, composition and function of objects. For example, these analyses can include histological tissue studies for disease detection, fluorescence imaging or stain-free analysis of cellular components, etc. In chemical analysis, imaging spectrometers can, for example, capture spectral features to identify molecular compounds. Imaging analysis can improve the ability to accurately diagnose diseases, study cellular interactions and advance medical research, helping to improve treatments and deepen the understanding of biological processes.

[0003] The typical problem with these analyses is that a large amount of data must be analyzed to obtain insightful results. This can take a lot of time and / or require a lot of IT resources. Therefore, there is a need to improve them. Summary of the invention

[0004] It is an object of the present disclosure to improve image-based analysis.

[0005] This object is solved by the disclosed embodiments. In particular, these embodiments are defined by the subject matter of the independent claims. The dependent claims provide information on other embodiments. Various aspects and embodiments of these aspects are also disclosed in the Summary of the Invention and the Detailed Description of the Invention, which provide other features and advantages.

[0006] A first aspect of the present disclosure relates to an image analysis method, comprising the steps of:

[0007] - Identify the first set of requirements;

[0008] - acquiring a plurality of images of the sample from an imaging device, in particular a microscope, according to the first set of requirements;

[0009] - obtaining a second set of requirements; and

[0010] - According to the second set of requirements, analyze the samples in the plurality of images.

[0011] The first set of requirements may be obtained based on user input and / or automatically obtained. In particular, the first set of requirements may be configured based on a specific analysis type and / or a specific device type to be used during the analysis. The first set of requirements may also be received and / or obtained from other devices, in particular from diagnostic / clinical / medical devices and / or communication devices (such as the cloud). The first set of requirements relates to the execution of image acquisition. In other words, the first set of requirements relates to the acquisition of information on which subsequent analysis is based. The first set of requirements may include one or more requirements.

[0012] Diagnostic and / or medical images may be acquired by various imaging modalities, such as one or more of the following image types:

[0013] - photographic images, for example for light microscopy, phase contrast microscopy or fluorescence microscopy;

[0014] - X-ray images;

[0015] -Computed tomography (CT) images;

[0016] - Magnetic resonance imaging (MRI) images; and

[0017] -Ultrasound, positron emission tomography (PET) technology. Image acquisition involves capturing images using an appropriate method and ensuring appropriate image quality.

[0018] The second set of requirements may be obtained by one or more of the methods described for the first set of requirements. The second set of requirements may include one or more requirements. The second set of requirements relates to the analysis itself and may include one or more requirements in the first set of requirements. Alternatively, the second set of requirements may include requirements different from the first set of requirements.

[0019] The analysis of multiple images according to the second set of requirements may include one or more analyses. Such analysis may include a frame-by-frame analysis, where each image is analyzed independently. Thus, image analysis techniques such as segmentation, feature extraction, and classification may be applied to each individual image. The results of the analysis of each image may be analyzed individually or combined to obtain meaningful insights.

[0020] Additionally or alternatively, the analysis of the plurality of images according to the second set of requirements may include temporal analysis. Thus, the image stream provides temporal information for analyzing changes and / or patterns over time. Temporal analysis may include detecting motion, tracking objects, identifying temporal patterns, or quantifying temporal dynamics by comparing consecutive frames. For this purpose, techniques such as optical flow, background subtraction, and tracking algorithms may be used.

[0021] Additionally or alternatively, the analysis of the plurality of images according to the second set of requirements may include event detection analysis. Event detection analysis may focus on identifying specific events and / or activities in the acquired images. This may include identifying predefined actions, unusual events, and / or significant changes in the scene. Machine learning techniques such as recursive neural networks and / or hidden Markov models may be used for event detection in image sequences.

[0022] Additionally or alternatively, the analysis of the plurality of images according to the second set of requirements may include target tracking. Target tracking may be used to track and locate a particular object of interest across the plurality of images. Target tracking may include associating an object detected in one frame with its corresponding instance in a subsequent frame in order to analyze the trajectory, velocity, interaction or behavior of the object over time. The tracking algorithm may be based on methods such as Kalman filtering, particle filtering and / or data association techniques.

[0023] Additionally or alternatively, the analysis of the plurality of images according to the second set of requirements may include object tracking. Motion analysis in an image stream (i.e., acquired images) may provide insights into dynamic events and activities. Techniques such as optical flow analysis, background subtraction, and / or motion energy analysis may be used to extract motion information from the image stream. This information may be further analyzed to detect anomalies, characterize motion patterns, or explain object dynamics.

[0024] Additionally or alternatively, the analysis of the plurality of images according to the second set of requirements may include change detection. Analysis based on change detection may focus on identifying differences and / or changes between consecutive frames. This may be used to detect and highlight regions of interest that have undergone changes, such as new objects, the appearance / disappearance of objects, or changes in scenes. Techniques such as image differencing, image registration, or pixel-level comparison may be used for change detection.

[0025] Additionally or alternatively, the analysis of the plurality of images according to the second set of requirements may include analysis based on fusion of the plurality of images. Rather than analyzing the individual images separately, the plurality of images may be fused or combined to extract higher level information. Such analysis may include image averaging, background modeling, and / or cross-image feature aggregation.

[0026] Additionally or alternatively, the analysis of the plurality of images according to the second set of requirements may include real-time analysis. Real-time analysis may include designing algorithms capable of processing and analyzing images within a given time limit and / or utilizing parallel processing of different analyses.

[0027] According to the method of the first aspect, resources used for analysis can be reduced.

[0028] A first aspect of the present disclosure relates to an image analysis method, wherein a first set of requirements includes one or more requirements related to content of one or more images.

[0029] Image content may refer to information, features and / or visual elements within an image. For example, it may include objects, patterns, textures, colors, shapes, frequencies, functions and / or structures visible in the image. Content-based requirements may relate to image features related to the environment in which the image or images were captured. In the case of microscopy images, this may be the morphological, (bio)chemical and / or molecular biological context. One or more content-related parameters may be set as a first requirement, particularly depending on the subsequent analysis.

[0030] Content-based requirements can include segmentation. This allows one or more images to be divided into meaningful regions or objects of interest. For example, segmentation can be used to identify and separate specific anatomical and / or morphological structures, lesions, organelles, or anomalies. A variety of algorithms such as thresholding, region growing, edge detection, and clustering can be used for segmentation.

[0031] Content-based requirements may include feature extraction. For example, once regions of interest have been identified through segmentation, one or more relevant features may be extracted from these regions. Features may include geometric properties (e.g., size, shape, volume), texture features, intensity statistics, or other quantitative metrics describing structures, regions, and / or anomalies within the image. Content-based requirements may include subcellular structures, such as mitochondria, nuclei, spindles, and / or other tissue structures.

[0032] Content-based requirements may include feature extraction. The extracted features may be used to classify and / or identify specific patterns, structures, or abnormalities in diagnostic and / or medical images. Feature extraction may be performed using machine learning algorithms such as convolutional neural networks. In addition, classification algorithms such as machine learning techniques (e.g., support vector machines, random forests, convolutional neural networks) may be used to distinguish between different regions and / or identify specific conditions.

[0033] Content-based requirements may include quantitative analysis. Quantitative analysis may involve the measurement and / or numerical evaluation of specific parameters and / or features within an image. This may include measurement of size, volume, density, intensity, and / or other quantitative indicators that may aid in diagnosis, monitoring, or treatment planning.

[0034] Content-based requests enable imaging devices to acquire only the images needed for a specific analysis. This can reduce stress on the probe, such as phototoxicity, and save resources.

[0035] A first aspect of the present disclosure relates to an image analysis method, wherein the first set of requirements includes one or more of the following requirements:

[0036] - spatial location;

[0037] - Identification of the sample-related carrier;

[0038] - time-related parameters;

[0039] - Wavelength-dependent parameters;

[0040] - Poor strength;

[0041] - motion parameters;

[0042] -pattern.

[0043] Spatial position can be related to the x, y, z translation coordinates of the subject object being analyzed (i.e., the probe in the microscope) in three-dimensional space. These coordinates can represent the lateral (horizontal) position of the object along the x-axis and y-axis, and the axial (vertical) position along the z-axis. Therefore, the image field of the subject object can be accurately determined. By knowing the position of the subject object, it can be ensured that the desired area or feature is being inspected.

[0044] In microscopy, for example, the spatial position of a sample can be associated with a stage position, which determines the position of the probe relative to a fixed reference point of the microscope. The spatial position of the subject object being analyzed can be associated with the focal plane of the microscope. The focal plane is a specific plane within the subject object (also called a "specimen" or "sample") that is sharp and other planes may appear blurry. Accurately positioning the sample within the focal plane is critical to capturing clear, detailed images of the sample and its surroundings. In some cases, multiple images (such as microscope images) can be acquired and stitched together to create a larger composite image of the sample or sample. The spatial position of the sample can be associated with one or more overlapping regions between these images to ensure that sample data from different frames are accurately aligned and / or seamlessly integrated.

[0045] Depending on the specific application or imaging technique, microscopy samples can be associated with different coordinate systems. For example, in fluorescence microscopy, samples can be associated with the coordinate system of the fluorescent channels and / or specific fluorophores used for labeling. Acquiring images of samples based on their spatial position in the associated coordinate system facilitates later interpretation and analysis of the acquired data.

[0046] When performing multimodal or multispectral imaging, the spatial position of the sample can be associated with the registration of different image modalities and / or channels. Image registration involves aligning images acquired from different sources and / or imaging modalities to ensure accurate spatial correspondence. The position of the probe can be used for registration purposes to allow the fusion or superposition of different imaging data.

[0047] The first set of requirements may be time-related parameters. A time-related parameter may be exposure time. Exposure time refers to the length of time that the sensor or detector of an image device is exposed to light during image capture. Proper exposure time is critical to achieve the best image quality and avoid overexposure or underexposure. Longer exposure times capture more light and increase image brightness.

[0048] A time-related parameter may be the frame rate. The frame rate determines the number of images captured per unit time. It represents the temporal resolution of the image sequence and is usually expressed in frames per second (fps). A higher frame rate can capture faster moving parts of the sample and / or more temporal details of dynamic processes.

[0049] Time-dependent parameters can be related to time-lapse imaging: Time-lapse imaging can involve capturing a series of images at specific time intervals over a longer period of time. In this way, time-dependent processes, changes or interactions related to the subject being analyzed can be observed. The length of time between adjacent images in the time-lapse sequence can be predetermined based on the desired temporal resolution and / or the dynamics of the phenomenon being observed.

[0050] The time-related parameter may be related to the speed and / or rate at which the imaging device captures and processes images. Faster imaging speeds are critical for real-time and / or dynamic applications.

[0051] A time-related parameter can be the exposure dwell time. For example, in scanning microscopy techniques such as confocal microscopy or scanning electron microscopy (SEM), the probe is scanned by a beam. The exposure dwell time is the length of time the beam stays at each pixel or point during the scanning process. The dwell time affects the signal-to-noise ratio, resolution, and image quality.

[0052] A time-related parameter can be acquisition time. Acquisition time refers to the total length of time required to capture a single image and / or an entire image sequence, and can include all the processes of image acquisition, including focusing, exposure, signal reading, and data transmission. Minimizing acquisition time is very important for capturing fast events or reducing the impact of motion artifacts.

[0053] The first set of requirements may be frequency. The frequency may relate to the frequency used by the imaging device to acquire the image. For example, the frequency for near-field microscopy may include optical frequency and / or infrared frequency.

[0054] The first set of requirements may be intensity differences. For example, in a microscope image, intensity differences may relate to changes and / or differences in brightness and / or intensity levels of different regions and / or features within the image. Intensity differences may represent contrast and / or distinctions between regions of different light intensities. Intensity differences in microscope images may be caused by a variety of factors, such as absorption or transmission properties and / or refractive index changes of the sample. In addition, intensity differences may also be related to fluorescence emission in fluorescence microscopy. Fluorescence emission may be related to the fluorescence signal of specific regions and / or structures within the sample.

[0055] The first set of requirements may be related to motion parameters. Motion parameters may be related to dynamic processes of the sample, the imaging device, and / or the imaging device environment, which may be revealed based on the acquired images. For example, in live cell imaging, the motion of cells, organelles, and / or other cellular components may be indicative of cell activity, migration, or response to stimuli. Motion-related parameters may be related to cell movement, flow analysis, particle tracking, tissue morphology, and time-lapse imaging. Motion parameters may also be related to mechanical properties of the sample, such as stiffness, elasticity, or surface deformation.

[0056] The first set of requirements may be related to a pattern, in particular a predetermined pattern. Analyzing the pattern of a sample by an imaging device may provide information on the structure, organization and / or characteristics of the sample.

[0057] Patterns can include regular structures such as crystals, lattice arrangements or repeating patterns, as well as irregular or randomly distributed features. The presence of symmetrical patterns and / or symmetry breaking (deviations of the pattern from a certain symmetry) can provide insight into the development, growth or response of the sample to external factors. Crystallographic related parameters used for polarizing microscopy and X-ray diffraction, etc., can reveal lattice and crystallographic properties. Additionally or alternatively, patterns can be related to cellular arrangements, surface textures, fractals and self-similarity, spatial distribution (especially in fluorescence microscopy) and / or biomolecular patterns, for example, the biomolecular patterns are used to identify biomolecules such as proteins, DNA and RNA within cells and tissues.

[0058] A first aspect of the present disclosure relates to an image analysis method, wherein a first set of requirements includes one or more requirements related to the form of one or more images.

[0059] One or more form-related parameters may be set as first requirements, particularly depending on the subsequent analysis. The acquired image may undergo a preprocessing step to improve image quality, reduce noise, correct artifacts and normalize image features. Preprocessing techniques may include filtering, image registration, denoising and image normalization. One or more of these preprocessing features may also be set as first requirements.

[0060] A first aspect of the present disclosure relates to an image analysis method, wherein the first set of requirements includes one or more of the following requirements:

[0061] - Image size;

[0062] - Image aperture; and

[0063] -Image resolution.

[0064] The image size may be related to the image form and / or the x, y, z dimensions of the image. The image size may also be related to focus, zoom, tilt and rotation. Additionally or alternatively, the first requirement related to the image size may be related to the detected region of interest (such as a structure) and the associated environment (especially the environment immediately adjacent to the region of interest).

[0065] Contrast refers to the ability of a microscope to distinguish a specimen from its background and / or surrounding material.

[0066] Aperture can refer to aperture size and / or depth of field. Depth of field refers to the thickness of the plane of focus. A larger depth of field allows a larger portion of the specimen to be in focus at the same time, providing a more comprehensive view. Aperture can also relate to illumination.

[0067] The first set of requirements, related to image resolution, can regulate the level of detail that can be observed. Image resolution can also relate to the magnification of the image. The higher the magnification, the more detailed the characteristics of the sample can be observed. Image resolution can also relate to temporal resolution, which is related to the speed at which the imaging device acquires the image.

[0068] For the latter analysis, the overall quality of the image is very important in the analysis. Therefore, the first set of requirements may also include parameters related to sharpness, clarity, color accuracy and / or the absence of artifacts and / or distortion.

[0069] A first aspect of the present disclosure relates to an image analysis method, wherein acquiring a plurality of images comprises:

[0070] -Storing subsets of images provided by the imaging system.

[0071] By storing reduced image streams, less storage and faster data processing for later analysis can be achieved.

[0072] Storing images from an imaging device, in particular a microscope, may involve managing and / or saving image data for later analysis. The storage may be associated with a file format, such as a TIFF and / or RAW format. The storage may further depend on a predetermined resolution of one or more images, in particular the images of the reduced image stream may have different file and / or resolution formats. Additionally or alternatively, the storage may also include metadata for one or more images in the reduced image stream. The metadata provides context information that aids in image interpretation and analysis, in particular relevant information arising during the image capture process. In particular, one or more storage parameters may be stored based on the first set of requirements and / or the second set of requirements.

[0073] A first aspect of the present disclosure relates to an image analysis method, wherein acquiring a plurality of images comprises the steps of: - controlling an imaging system for capturing images.

[0074] Controlling an imaging system, especially an imaging system of a microscope, may include controlling various parameters, especially parameters included in the first set of requirements and / or parameters related thereto. Control of the imaging system may include control of objective lens selection, control of one or more focus adjustment parameters, control of illumination settings, control of shutter speed and / or exposure time. Additionally or alternatively, control of the imaging device may also include control of filter selection. In particular, in fluorescence microscopy, specific fluorescence wavelengths may be isolated and / or background noise may be minimized by selecting specific filters. Control of the imaging system may also include controlling the image acquisition process, especially triggering image capture, adjusting the frame rate of time-lapse imaging, and / or managing the image acquisition process of the total number of frames to be collected. In three-dimensional imaging, Z stack reconstruction may be achieved by controlling the microscope to capture a series of images at different focal planes. For fluorescence microscopy and multicolor imaging, a system for capturing multiple channels or spectral bands may be controlled by controlling the imaging system.

[0075] A first aspect of the present disclosure relates to an image analysis method, wherein acquiring an image includes acquiring one or more images triggered by an event.

[0076] An event that may trigger the step "acquiring a plurality of images" is associated with one or more of the first set of requirements. In particular, event-triggered image acquisition may acquire images in a timely and gap-free manner. Event-triggered acquisition of one or more images may result in the selection and storage of specific images from a stream of images captured by an imaging device. Additionally or alternatively, event-triggered acquisition of one or more images may also result in the capture of images by activating an imaging device. Thus, images may only be acquired if the requirements are met.

[0077] Events can trigger automatic time-lapse imaging, thereby capturing images at predetermined time intervals. This can be used to observe dynamic processes over long periods of time, such as cell growth or movement. Events can further trigger image acquisition when specific events occur, especially events related to the sample to be analyzed, such as when a certain fluorescence intensity threshold is reached in a live cell imaging experiment. Additionally or alternatively, events can further trigger image acquisition when external events occur (such as changes in temperature, pressure or other environmental conditions). Events can be further defined by image acquisition protocols. Based on the analysis results, in particular based on the first and / or second set of requirements, predefined protocols for capturing specific images can be formulated and the imaging device can be programmed to follow these protocols to ensure consistency and repeatability of image acquisition.

[0078] A first aspect of the present disclosure relates to an image analysis method, wherein acquiring a plurality of images is performed based on a machine learning algorithm.

[0079] Different machine learning algorithms, such as convolutional neural networks (CNNs), can be used, especially when performing image classification, target detection and / or segmentation tasks. In microscopy, CNNs can be used to identify and classify cells, organelles and / or other structures of interest within an image. Additionally or alternatively, semantic segmentation can also be used to divide each pixel in an image into specific categories and / or regions of interest. Convolutional neural networks with architectures such as U-Net or DeepLab can be used to perform semantic segmentation tasks in microscopy. Additionally or alternatively, instance segmentation can be used to identify and / or distinguish individual instances of objects within an image. Region-based convolutional neural networks and / or similar architectures can be used to perform instance segmentation tasks in microscopy. Additionally or alternatively, target detection algorithms can locate and classify multiple objects of interest within an image. In addition, recursive neural networks (RNNs), RNNs and their variants, such as long short-term memory (LSTM) networks, can also be used to analyze sequential data based on multiple images acquired. They can be applied to time series microscopy data and / or videos to capture temporal dependencies in cell behavior or other dynamic processes. Random forests and / or support vector machines (SVM) may be used to analyze the image, in particular to identify different regions in the image.

[0080] A first aspect of the present disclosure relates to an image analysis method, wherein a machine learning algorithm is executed based on a first set of requirements and / or a second set of requirements.

[0081] The result of the machine learning algorithm can meet the first set of requirements. At this point, the acquisition of one or more images can be automatically triggered. For example, based on CNN or SVN, a pre-defined state of the sample can be identified in the "pre-image" of the sample continuously taken by the imaging device. Based on this event, the imaging device can be triggered to capture a "full image" (the pre-image and the full image may differ under lighting conditions). Alternatively, for example, if the full image is captured continuously / continuously, image selection and image storage can be triggered. In the above two schemes, only complete images that meet the first set of requirements monitored by the machine learning algorithm can be stored. Therefore, the machine learning algorithm can be executed under soft real-time or hard real-time conditions.

[0082] A first aspect of the present disclosure relates to an image analysis method, wherein the machine learning algorithm comprises a neural network, in particular a convolutional neural network.

[0083] Based on a neural network, one or more of the first set of requirements can be monitored. The neural network can include a standard neural network, such as VGG and / or ResNET. VGG is a convolutional neural network architecture characterized by the use of small (3×3) convolutional filters and deep stacking of convolutional layers. ResNET is a deep convolutional neural network architecture that uses residual connections to train ultra-deep neural networks, which can be up to hundreds of layers deep. In particular, convolutional neural networks can be used for efficient image processing. To determine events, states and / or structures within a sample, a fully connected single-layer or multi-layer recurrent neural network can be used. The recurrent neural network can be a feedforward neural network, such as a multilayer perceptron. The recurrent neuron network can be configured in particular to obtain its input from a convolutional neural network. Therefore, the segmentation of the obtained one or more images can be calculated with particularly high accuracy.

[0084] A first aspect of the present disclosure relates to an image analysis method, wherein the step of acquiring a plurality of images from an imaging device comprises:

[0085] - acquiring a plurality of images from a first device; and

[0086] - Acquire a plurality of images from one or more other devices.

[0087] The process of acquiring images from the first device and one or more other devices can be performed simultaneously and / or sequentially. For example, the first device and the other devices can be part of a distributed microscopy system, connected via a network. The microscopes of the distributed system can be located in the same location or in different locations, and they work in coordination to jointly capture images and / or data of the same and / or different samples. This setup can quickly capture relevant images for subsequent analysis.

[0088] A first aspect of the present disclosure relates to an image analysis method, wherein while acquiring a plurality of images of a sample, the method comprises:

[0089] - Controlled influence on the sample, e.g. providing liquid and / or mechanical pressure.

[0090] The imaging device and / or the additional device may produce an effect on the sample. The additional device may be a liquid handling device or the like for adding liquid to the sample when a predefined condition is met. For example, the condition may be the first or second set of requirements. Additionally or alternatively, the effect may be a mechanical / fluid dynamic effect, such as pressure; a biochemical effect, such as the addition of an agent; and / or a molecular effect, such as a DNA probe inserted into the specimen.

[0091] A second aspect of the present disclosure relates to a diagnostic and / or medical device, in particular a microscope, for:

[0092] - performing the method according to the first aspect and / or any other embodiment; or

[0093] - Receiving control information in the method according to the first aspect and / or any other embodiment.

[0094] Such a device may be an imaging device. Additionally or alternatively, according to the above embodiment of the first aspect, such a device may be an additional device configured to exert a predetermined influence on the sample. Additionally or alternatively, such a device may also be a computing device for receiving a sample image from the imaging device and / or controlling the imaging device and / or an additional device configured to exert an influence on the sample.

[0095] A third aspect of the present disclosure relates to a graphical user interface (GUI), in particular a GUI for a diagnostic and / or medical device, configured as follows:

[0096] - define content-related requirements and / or format-related requirements for image acquisition;

[0097] -a) display one or more images according to defined requirements;

[0098] -Access to the device described in the above aspects.

[0099] The GUI can be used to define one or more first set of requirements for acquiring images of one or more samples. For example, a user can provide a region of interest; the imaging device is controlled based on the region of interest and focuses on the region in one or more images. After analysis, the results are visualized and presented in a manner that is convenient for medical professionals to interpret. Visualization techniques can include two-dimensional and three-dimensional rendering, multi-planar reconstruction, surface rendering, or other methods that can effectively display the analyzed data.

[0100] In some cases, the GUI can visualize medical and / or diagnostic decision support systems to help healthcare professionals make accurate and informed diagnostic or treatment decisions. These systems can combine image analysis algorithms, clinical guidelines, patient data, and / or other relevant information to provide advice or assistance to medical practitioners.

[0101] The GUI allows visualization of correlation functions. Thus, imaging information (e.g. microscopy data, morphological data) can be linked to (bio)chemical data (e.g. proteins) and / or molecular biology data (e.g. information on RNA, DNA). Via the GUI, the user can also select which information should be associated with the image information for analysis purposes. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] The following embodiments (partially shown in the accompanying drawings) have further advantages and features. These embodiments are not all shown to scale in the accompanying drawings. The dimensions of various features may be enlarged or reduced, especially for clarity of description. For this reason, at least some of these drawings are schematic diagrams.

[0103] Figure 1 The setup and workflow of one embodiment of the present disclosure for analyzing samples is shown.

[0104] Figure 2 A flow chart of a method according to an embodiment of the present disclosure is shown.

[0105] Figure 3 A convolutional neural network configured for an embodiment of the present disclosure is shown.

[0106] Figure 4 A microscope system that may be used with embodiments of the present disclosure is shown.

[0107] In the following description, please refer to the accompanying drawings which form a part of the present disclosure and illustrate specific aspects that can be understood. The same reference symbols indicate features that are the same or at least similar in function or structure. DETAILED DESCRIPTION

[0108] In general, the disclosure of a described method also applies to a corresponding device (or apparatus) for performing the method or a corresponding system consisting of one or more devices, and vice versa. For example, if a specific method step is described, the corresponding device may include characteristics for performing the method step, even if the characteristic is not explicitly described or represented in the figure. On the other hand, for example, if a specific device is described in terms of functional units, the corresponding method may include one or more steps to perform the described function, even if these steps are not explicitly described or represented in the figure. Similarly, a system may have corresponding device characteristics or characteristics for performing specific method steps. Unless otherwise explicitly stated, the features of the various exemplary aspects and embodiments described above or below may be combined.

[0109] Figure 1 The setup and workflow of an embodiment 100 of the present disclosure is shown. The first step in the experiment and its data generation is to set up and prepare for analysis by defining a first set of requirements 102, which includes, among other things, the structure of the region of interest in the image that can be captured by the microscope used in the experiment. A machine learning-based detection (and selection) algorithm is trained 104 according to the first set of requirements. Optionally, a pre-trained algorithm can be used. Such an algorithm is used to determine the type of data (structure, event, intensity, color, motion, etc.) required to perform subsequent analysis. The machine learning algorithm automatically sets the necessary acquisition settings to deliver images according to the first set of requirements.

[0110] For example, a machine learning-based detection (and selection) algorithm could be based on Figure 3 A convolutional neural network (CNN) is shown. The trained CNN can identify regions of interest in images acquired by the microscope. The CNN can be trained to enable rapid recognition (in real time) so that changes in the sample can be recognized shortly after they occur. The machine learning algorithm is then connected to the imaging device 106 so that it can control image acquisition according to the first set of requirements. The first three steps 102, 104, 106 constitute the pre-analysis configuration / setup 108.

[0111] Afterwards, the configured image acquisition and data analysis 110 can be performed. Based on the first set of requirements, the CNN controls the imaging device to only capture images that meet the first set of requirements. Image capture is completed in step 112. All other images are not captured. This allows the generated image stream to be focused on images that are valuable for subsequent analysis. In other words, through the conditional image acquisition process 112, data reduction 114 is achieved.

[0112] Based on the reduced image data 112, 114, an analysis 116 is performed. The analysis is performed based on a second set of requirements that is different from the first set of requirements upon which the image selection was based.

[0113] A graphical user interface (GUI) is used to guide the user through the analysis workflow 118. The user can define a region of interest, for example based on existing, suggested examples, and then use this information throughout the CNN analysis process.

[0114] The analysis settings can be adjusted (automatically and / or by the user through the GUI) according to the images to be acquired. Settings such as signal-to-noise ratio, frame acquisition, number and "thickness" of Z steps (perpendicular to the focal plane), imaging speed, etc. can be defined by the user and / or automatically adjusted, especially according to the first set of requirements. In addition, parameters such as camera settings, delay settings and / or position settings can also be defined.

[0115] Image acquisition can be controlled using different machine learning based detection algorithms. Different machine learning based detection algorithms can be trained based on different first requirements. For example, one detection algorithm can be trained to recognize cell divisions and another detection algorithm can be trained to recognize certain cell structures. The appropriate detection algorithm can be automatically selected based on the first set of requirements.

[0116] Additionally or alternatively, the user may also select a machine learning detection algorithm through the GUI. This is particularly useful if there is more than one detection algorithm suitable for a particular analysis and / or first set of requirements. Thus, the GUI may be used to propose one or more detection algorithms to the user. The GUI may also be used to further train the detection / selection algorithm. Additionally or alternatively, the GUI may also be used to display relevant information about the selected image acquisition method (i.e., the selected detection / selection method).

[0117] The first and / or second set of requirements can be visualized based on the reduced set of images analyzed, for example, a 6D visualization related to parameters such as xyz coordinates of the region of interest, changes over time, wavelength distribution of the region of interest, intensity differences in the region of interest and / or the position / ID of the sample or sample carrier (such as different wells of a multi-well plate). External biochemical and / or molecular biological (DNA, RNA, omics data) information can also be integrated into the analysis results and associated with the analysis results. In particular, a spatial context can be provided in which the first set of requirements, the second set of requirements and / or additional information can be intuitively presented to the user.

[0118] According to the disclosed method, fewer images need to be taken for analysis, reducing phototoxicity and bleaching. This can improve the quality of the acquired images and ultimately improve the quality of image-based analysis. Unlike current imaging / analysis systems, the workflow and user experience provided by this embodiment starts with analysis rather than (unconditional) imaging.

[0119] An advantage of this embodiment is that it reduces the large amount of data that may be generated by standard imaging acquisition methods. Reducing the amount of data to be analyzed can also reduce the required processing resources and analysis time, especially in long-term time-delay experiments and / or high-content screening experiments. The described embodiment can reduce the amount of data to be stored in main storage and / or long-term storage. In addition, this embodiment can also release the space and processing load of the GPU and / or CPU of the computer that performs the method described in this embodiment. Feedback microscopy, autonomous imaging and analysis, and three-dimensional visualization can be performed more efficiently and smoothly, and allow imaging information to be associated with data from other technologies, for example, the spatial association of structure and function.

[0120] The embodiments described in the present disclosure can be used for various types of experiments, such as experiments related to: detection and analysis of structures / sample types / events / changes over time; detection and analysis of intracellular structures (such as mitotic spindles, mitochondria, nuclei, membranes, vesicles, cytoskeleton); detection and analysis of two-dimensional cells (such as cell types, expression patterns, "positive / negative"); detection and analysis of three-dimensional cell cultures (such as spheroids, organoids, morphology, expression, drug testing); detection and analysis of tissue sections (such as functional / morphological regions, patterns); detection and analysis of organisms (such as organs, protozoa, multicellular organisms, host-pathogens, model organisms); samples on slides, dishes, and well plates.

[0121] The embodiments described in the present disclosure can be used for various types of analysis and / or detection methods, such as analysis / detection methods related to: morphology / pathomorphology, rare events, cell cycle, cell division, movement / tracking / motion, intensity / change / increase / decrease / quantification, wavelength / color, structure counting, structure interaction, spheroid / organoid formation, gene expression, protein distribution, proliferation, development, apoptosis, live / dead analysis, effects of drugs / compounds.

[0122] The embodiments described in the present disclosure can be used for various types of applications, such as those related to intracellular transport mechanisms, molecular interactions and dynamics, signal transduction, regenerative medicine, stem cells, genomics, proteomics, cancer research and diagnosis, immunology, embryogenesis, neural development, organ development.

[0123] Figure 2A flow chart of a method 200 according to an embodiment of the present disclosure is shown. In a first step 210, a first set of requirements is obtained. The first set of requirements defines the content and form of the image to be captured. By only capturing images with predetermined content / form, analysis based on the captured images can use fewer resources and achieve better results. The first step requirements can be provided by a database, depending on the type of analysis to be performed. Additionally or alternatively, the user can also provide the first set of requirements through a graphical user interface (GUI).

[0124] In a second step 220, a second set of requirements is obtained. The second set of requirements defines the requirements according to which the analysis should be performed. In many cases, the first set of requirements is different from the second set of requirements. However, it is also possible that at least some of the requirements in the first set of requirements are included in the second set of requirements. The second step requirements can be provided by a database, depending on the type of analysis to be performed. Additionally or alternatively, the user can provide the second set of requirements via a GUI. The second step 220 must be performed (at least partially) before the analysis begins. The second step can be performed after, before or in parallel with the first step 210 and / or the third step 230.

[0125] In a third step 230, the imaging device is controlled to capture images that meet one or more, in particular all, of the first set of requirements. The imaging device may be a microscope or other imaging device for observing samples, specimens and / or other objects for subsequent analysis.

[0126] In a fourth step 240, the acquired image is analyzed according to the second set of requirements. This step can be initiated after the first image is captured and provided for analysis. Thereafter, the third step 230 and the fourth step 240 can be performed simultaneously unless sufficient image information has been acquired.

[0127] In a fifth step 250, the analyzed data is visualized for the user via a GUI. Thus, in particular, the results of the analysis are illustrated together with information related to one or more requirements of the first and / or second set of requirements and / or additional information. The GUI may be present throughout the entire process.

[0128] Figure 3 A neural network for a method according to an embodiment of the present disclosure is shown. The neural network 300 is used to determine a region of interest of an image 301 of a sample where cell division occurs. The neural network can identify not only images 301 containing the region of interest, but also images that do not contain the region of interest. The latter will not be captured / saved as images for analysis.

[0129] The neural network 300 may be a convolutional neural network of the standard type ("VGG16"). The entry layer includes two convolutional layers 302a of size 224×224×64. The first hidden layer includes a maximum pooling layer 304a, which samples the initial image using a 3×3 receptive field with a minimum sampling size of 112×112×128. The first hidden layer also includes two convolutional layers 302b. The second hidden layer performs downsampling to 56×56×256 via a maximum pooling layer 304b, followed by three convolutional layers 302c. The third hidden layer performs downsampling to 28×28×512 via a maximum pooling layer 304c, followed by three convolutional layers 302d. The fourth hidden layer performs downsampling to 14×14×512 via a maximum pooling layer 304d, followed by three convolutional layers 302e. The final maximum pooling layer 304e performs further downsampling to 7×7×512. This layer is connected to a single 8×1×1 fully connected layer 306 for regression analysis.

[0130] The neural network 300 is trained on discrete frames of the video with respective analysis types as input. The training data may include labeled regions of interest where cell division occurs. The labels are collected from human annotations. During training, the neural network is optimized through supervised back-propagation to maximize the intersection of the predicted bounding boxes of the regions of interest for each video frame with the labels collected by humans.

[0131] Some embodiments relate to a microscope comprising: Figures 1 to 3 Alternatively, the microscope may be a system associated with one or more of Figures 1 to 3 Part of or connected to one or more related systems. Figure 4 A schematic diagram of a system 400 configured to perform the methods described herein is shown. System 400 includes a microscope 410 and a computer system 420. Microscope 410 is configured to capture images and is connected to computer system 420. Computer system 420 is configured to perform at least a portion of the methods described herein. Computer system 420 can be configured to execute a machine learning algorithm. Computer system 420 and microscope 410 can be separate entities, but can also be integrated in a common housing. Computer system 420 can be part of a central processing system of microscope 410, and / or computer system 420 can be part of a subcomponent of microscope 410, such as a sensor, actuator, camera, or lighting unit of microscope 410, etc.

[0132] The computer system 420 may be a local computer device (e.g., a personal computer, a laptop, a tablet computer, or a mobile phone) having one or more processors and one or more storage devices, or may be a distributed computer system (e.g., a cloud computing system having one or more processors and one or more storage devices, distributed in different locations, for example, at a local client and / or one or more remote server farms and / or data centers). The computer system 420 may include any circuit or combination of circuits. In one embodiment, the computer system 420 may include one or more processors, which may be of any type. As used herein, a processor may represent any type of computing circuit, such as, for example, a microscope or a microscope component (e.g., a camera) or any other type of processor or processing circuit, such as, but not limited to, a microprocessor, a microcontroller, a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a graphics processor, a digital signal processor (DSP), a multi-core processor, a field programmable gate array (FPGA). Other types of circuits that may be included in the computer system 420 may be custom circuits, application specific integrated circuits (ASICs), and the like, such as one or more circuits (e.g., communication circuits) for wireless devices (e.g., mobile phones, tablet computers, laptops, two-way radios, and similar electronic systems). The computer system 420 may include one or more storage devices, which may include one or more storage elements suitable for a particular application, such as main memory in the form of random access memory (RAM), one or more hard disk drives, and / or one or more drives for handling removable media, such as compact disks (CDs), flash memory cards, digital video disks (DVDs), etc. The computer system 420 may also include a display device, one or more speakers, and a keyboard and / or controller, which may include a mouse, a trackball, a touch screen, a voice recognition device, or any other device that allows a system user to input information to and receive information from the computer system 420.

[0133] Some or all of the method steps may be performed by (or using) a hardware device, such as a processor, a microprocessor, a programmable computer or an electronic circuit. In some embodiments, one or more of some of the most important method steps may be performed by such a device.

[0134] Depending on the specific implementation requirements, embodiments of the present invention may be implemented in hardware or software. It may be implemented using a non-temporary storage medium (e.g., a digital storage medium, such as a floppy disk, DVD, Blu-ray, CD, ROM, PROM, EPROM, EEPROM, or FLASH memory) in which electronically readable control signals are stored that cooperate with (or can cooperate with) a programmable computer system to perform the corresponding method. Therefore, the digital storage medium may be computer readable.

[0135] Some embodiments according to the invention comprise a data carrier carrying electronically readable control signals, which are capable of cooperating with a programmable computer system, such that one of the methods described herein is performed.

[0136] Generally, the embodiments of the present invention can be realized as a computer program product with a program code, when the computer program product runs on a computer, the program code is used to perform one of the methods. The program code can, for example, be stored on a machine-readable carrier.

[0137] Other embodiments comprise the computer program for performing one of the methods described herein, stored on a machine readable carrier.

[0138] In other words, therefore, one embodiment of the inventive method is a computer program having a program code for performing one of the methods described herein, when the computer program runs on a computer.

[0139] Therefore, another embodiment of the present invention is a storage medium (or a data carrier or a computer-readable medium) having stored thereon a computer program for performing one of the methods described herein when executed by a processor. The data carrier, the digital storage medium or the recorded medium is typically tangible and / or non-transitory. Another embodiment of the present invention is an apparatus as described herein, comprising a processor and a storage medium.

[0140] Therefore, another embodiment of the invention is a data stream or a sequence of signals representing the computer program for performing one of the methods described herein.For example, the data stream or the sequence of signals may be configured to be transmitted via a data communication connection (eg via the Internet).

[0141] A further embodiment comprises a processing means, for example a computer or a programmable logic device, configured to or adapted to perform one of the methods described herein.

[0142] A further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.

[0143] Another embodiment of the invention comprises an apparatus or a system configured to transfer (for example, electronically or optically) a computer program for performing one of the methods described herein to a receiver. The receiver may be, for example, a computer, a mobile device, a storage device, etc. The apparatus or the system may comprise, for example, a file server for transferring the computer program to the receiver.

[0144] In some embodiments, a programmable logic device (e.g., a field programmable gate array) can be used to perform some or all of the functions of the methods described herein. In some embodiments, a field programmable gate array can collaborate with a microprocessor to perform one of the methods described herein. Typically, these methods are preferably performed by any hardware device.

[0145] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items and may be abbreviated as " / ".

[0146] Although some aspects are described in the context of an apparatus, it is clear that these aspects also represent a description of the corresponding method, where a block or device corresponds to a method step or a feature of a method step. Similarly, aspects described in the context of a method step also represent a description of a corresponding block or item or feature of the corresponding apparatus.

[0147] Embodiments may be based on the use of machine learning models or machine learning algorithms. Machine learning may refer to algorithms and statistical models that can be used by computer systems that do not require the use of explicit instructions, but rely on models and inferences to perform specific tasks. For example, in machine learning, data transformations inferred from the analysis of historical and / or training data may be used instead of rule-based data transformations. For example, the content of an image may be analyzed using a machine learning model or using a machine learning algorithm. In order for a machine learning model to analyze the content of an image, a machine learning model may be trained using training images as input and training content information as output. By training a machine learning model using a large number of training images and / or training sequences (e.g., words or sentences) and associated training content information (e.g., labels or annotations), the machine learning model "learns" to recognize the content of the image, so the machine learning model can be used to recognize the content of the image not included in the training data. The same principle applies to other types of sensor data: by training a machine learning model using training sensor data and expected outputs, the machine learning model "learns" the conversion between sensor data and output, which can be used to provide outputs based on non-training sensor data provided to the machine learning model. The provided data (e.g., sensor data, metadata, and / or image data) may be preprocessed to obtain a feature vector, which is used as an input to the machine learning model.

[0148] The machine learning model can be trained using training input data. The above example uses a training method called "supervised learning". In supervised learning, the machine learning model is trained using multiple training samples, each of which may include multiple input data values ​​and multiple expected output values, i.e., each training sample is associated with an expected output value. By specifying the training samples and the expected output values, the machine learning model "learns" which output value to provide based on input samples similar to the samples provided during training. In addition to supervised learning, semi-supervised learning can also be used. In semi-supervised learning, some training samples lack corresponding expected output values. Supervised learning can be based on supervised learning algorithms (e.g., classification algorithms, regression algorithms, or similarity learning algorithms). When the output is limited to a limited set of values ​​(categorical variables), a classification algorithm can be used, that is, the input is classified into one of the limited set of values. When the output can have any numerical value (within a certain range), a regression algorithm can be used. Similarity learning algorithms can be similar to classification and regression algorithms, but are based on learning from examples using a similarity function that measures the similarity or degree of correlation between two objects. In addition to supervised or semi-supervised learning, unsupervised learning can also be used to train machine learning models. In unsupervised learning, (only) input data may be provided, and an unsupervised learning algorithm may be used to find structure in the input data (e.g., find commonalities in the data by grouping or clustering the input data). Clustering is the distribution of input data comprising multiple input values ​​into subsets (clusters) such that input values ​​within the same cluster are similar according to one or more (predefined) similarity criteria, but are dissimilar to input values ​​included in other clusters.

[0149] Reinforcement learning is the third group of machine learning algorithms. In other words, reinforcement learning can be used to train machine learning models. In reinforcement learning, one or more software actors (called "software agents") are trained to take actions in an environment. Based on the actions taken, a reward is calculated. Reinforcement learning is based on training one or more software agents to select actions that increase the cumulative reward, thereby making the software agent perform better on a given task (evidence is the increase in reward).

[0150] In addition, some techniques can be applied to some machine learning algorithms. For example, feature learning can be used. In other words, a machine learning model can be trained at least in part using feature learning, and / or a machine learning algorithm can include a feature learning component. Feature learning algorithms (which can be called representation learning algorithms) can retain the information in their input, but can also transform it in a useful way, usually as a preprocessing step before performing classification or prediction. For example, feature learning can be based on principal component analysis or cluster analysis.

[0151] In some examples, anomaly detection (i.e., outlier detection) can be used, the goal of which is to identify input values ​​that are sufficiently different from the majority of inputs or training data to be suspicious. In other words, a machine learning model can be trained at least in part using anomaly detection, and / or a machine learning algorithm can include an anomaly detection component.

[0152] In some examples, a machine learning algorithm can use a decision tree as a prediction model. In other words, a machine learning model can be based on a decision tree. In a decision tree, an observation about an item (e.g., a set of input values) can be represented by a branch of the decision tree, and an output value corresponding to the item can be represented by a leaf of the decision tree. A decision tree can support both discrete values ​​and continuous values ​​as output values. If discrete values ​​are used, the decision tree can be represented as a classification tree, and if continuous values ​​are used, the decision tree can be represented as a regression tree.

[0153] Association rules are another technique that can be used with machine learning algorithms. In other words, a machine learning model can be based on one or more association rules. Association rules are created by identifying relationships between variables in a large amount of data. Machine learning algorithms can identify and / or utilize one or more relationship rules that represent knowledge derived from the data. These rules can be used to store, manipulate, or apply knowledge.

[0154] Machine learning algorithms are often based on machine learning models. In other words, the term "machine learning algorithm" may refer to a set of instructions that can be used to create, train, or use a machine learning model. The term "machine learning model" may refer to a data structure and / or a rule set that represents the learned knowledge (e.g., based on the training performed by the machine learning algorithm). In an embodiment, the use of a machine learning algorithm may mean the use of an underlying machine learning model (or multiple underlying machine learning models). The use of a machine learning model may mean that the machine learning model and / or the data structure / rule set that is the machine learning model is trained by a machine learning algorithm.

[0155] For example, a machine learning model can be an artificial neural network (ANN). ANN is a system inspired by biological neural networks, such as neural networks in the retina or brain. ANN includes multiple interconnected nodes and multiple connections between nodes, so-called edges. There are usually three types of nodes: input nodes that receive input values, hidden nodes that are (only) connected to other nodes, and output nodes that provide output values. Each node can represent an artificial neuron. Each edge can transfer information from one node to another. The output of a node can be defined as a (non-linear) function of its input (e.g. the sum of its inputs). The input of a node can be used in a function based on the "weight" of the edge or the node providing the input. The weights of the nodes and / or edges can be adjusted during the learning process. In other words, the training of an artificial neural network can include adjusting the weights of the nodes and / or edges of the artificial neural network, i.e. achieving the desired output for a given input.

[0156] Alternatively, the machine learning model can be a support vector machine, a random forest model, or a gradient boosting model. A support vector machine (i.e., a support vector network) is a supervised learning model with an associated learning algorithm that can be used to analyze data (e.g., in classification or regression analysis). A support vector machine can be trained by providing an input with multiple training input values ​​belonging to one of two categories. A support vector machine can be trained to assign new input values ​​to one of two categories. Alternatively, the machine learning model can be a Bayesian network, which is a probabilistic directed acyclic graph model. A Bayesian network can use a directed acyclic graph to represent a set of random variables and their conditional dependencies. Alternatively, the machine learning model can be based on a genetic algorithm, which is a search algorithm and heuristic technique that mimics the natural selection process.

[0157] Reference numerals list

[0158] 100 Sample Analysis Method

[0159] 102 Limiting the first requirement

[0160] 104 Training Machine Learning Algorithms

[0161] 106 Connecting ML algorithms to selected imaging devices

[0162] 108 Configuration Analysis

[0163] 110 Image acquisition and analysis

[0164] 112 Selected collection

[0165] 114 Reduced Data

[0166] 116 Analysis of reduced data

[0167] 118 Visualizing the Results

[0168] 200 Method Flowchart

[0169] 210 Get the first set of requirements

[0170] 220 Get the second set of requirements

[0171] 230 Controlling imaging device

[0172] 240 Analyze images

[0173] 250 Visualizing Data

[0174] 300 CNNs for Image Analysis

[0175] 301 images

[0176] 302a Convolutional layer at the entrance layer

[0177] 302b Convolutional layer of the first hidden layer

[0178] 302c Convolutional layer of the second hidden layer

[0179] 302d Convolutional layer of the third hidden layer

[0180] 302e Convolutional layer of the fourth hidden layer

[0181] 304a The maximum pooling layer of the first hidden layer

[0182] 304b The second hidden layer's maximum pooling layer

[0183] 304c The third hidden layer's maximum pooling layer

[0184] 304d The fourth hidden layer's maximum pooling layer

[0185] 304e Final max pooling layer

[0186] 306 Fully connected layer

[0187] 400 Analysis System

[0188] 410 Microscope

[0189] 420 Computer

Claims

1. A computer-based method for analyzing a plurality of images, in particular for analyzing diagnostic and / or medical images, Includes steps: - determining a first set of requirements (102); - acquiring a plurality of images (112) of the sample from an imaging device, in particular a microscope, according to the first set of requirements; - obtain the second set of requirements; - Analyzing samples in the plurality of images according to a second set of requirements (116).

2. The method according to the preceding claim, in, The first set of requirements (102) includes one or more requirements related to the content of one or more images.

3. The method according to the preceding claim, in, The first set of requirements (102) includes one or more of the following requirements: - spatial location; - Identification of the sample-related carrier; - time-related parameters; - Wavelength-dependent parameters; - Poor strength; - motion parameters; and -pattern.

4. The method according to claim 1, in, The first set of requirements (102) includes one or more requirements related to the form of the one or more images.

5. The method according to the preceding claim, in, The first set of requirements (102) includes one or more of the following requirements: - Image size; - Image aperture; and -Image resolution.

6. A method according to any one of the preceding claims, in, Acquiring a plurality of images (112) comprises the steps of: -Storing subsets of images provided by the imaging system.

7. A method according to any one of the preceding claims, in, Acquiring a plurality of images (112) comprises the steps of: -Controlling an imaging system for capturing images.

8. The method according to any one of the two preceding claims, in, Acquiring a plurality of images (112) includes acquiring one or more images triggered by an event.

9. A method according to any one of the preceding claims, in, Acquiring a plurality of images (112) is performed based on a machine learning algorithm.

10. The method according to the preceding claim, in, The machine learning algorithm is executed based on the first set of requirements (102) and / or the second set of requirements.

11. The method according to any one of the two preceding claims, in, The machine learning algorithm comprises a neural network (300), in particular a convolutional neural network.

12. A method according to any one of the preceding claims, in, The step of acquiring a plurality of images from an imaging device comprises: - acquiring a plurality of images from a first device; and - Acquire a plurality of images from one or more other devices.

13. A method according to any one of the preceding claims, in, While acquiring a plurality of images of the sample, the method includes: - controlling the influence on the sample, for example providing liquid and / or mechanical pressure.

14. An imaging device (410), in particular a microscope, Constructed into: - performing a method (200) according to any of the preceding claims; or - Receiving control information according to a method (220) according to any of the preceding claims.

15. A graphical user interface, Constructed into: - defining content-related requirements and / or format-related requirements for image acquisition (102); - displaying one or more images (118) according to defined requirements; and - Access to a device (106) as claimed in the preceding claim.

Images