In a disclosed example, a computer-implemented method includes storing image data that includes an input image of a blood sample within a blood monitoring device. The method also includes generating, by a machine learning model, a segmentation mask that assigns pixels in the input image to one of a plurality of classes, which correlate to respective known biophysical properties of blood cells. The method also includes extracting cell images from the input image based on the segmentation mask, in which each extracted cell image includes a respective cluster of the pixels assigned to a respective one of the plurality of classes.

The present disclosure provides methods of preparing a biological specimen for imaging analysis, comprising fixing and clearing the biological specimen and subsequently analyzing the cleared biological specimen using microscopy. Also included are methods of quantifying cells, for example, active populations of cells in response to a stimulant. The present disclosure also provides devices for practicing the described methods. A flow-assisted clearing device provides rapid clearing of hydrogel-embedded biological specimens without the need of specialized equipment such as electrophoresis or perfusion devices.

A computer-implemented method is provided for analyzing videos of a living system captured with microscopic imaging. The method can include obtaining a base dataset including one or more videos captured with microscopic imaging with at least one of the one or more videos including a cellular event, and cropping out, from the base dataset, sub-videos including one or more objects of interest that may be involved in the cellular event. An artificial neural network (ANN) model can be trained using the plurality of selected sub-videos as training data, to perform unsupervised video alignment, a query sub-video can be aligned using the trained ANN model, and a determination can be made whether or not the query sub-video includes the cellular event.

In various examples, high beam control for vehicles may be automated using a deep neural network (DNN) that processes sensor data received from vehicle sensors. The DNN may process the sensor data to output pixel-level semantic segmentation masks in order to differentiate actionable objects (e.g., vehicles with front or back lights lit, bicyclists, or pedestrians) from other objects (e.g., parked vehicles). Resulting segmentation masks output by the DNN(s), when combined with one or more post processing steps, may be used to generate masks for automated high beam on/off activation and/or dimming or shading—thereby providing additional illumination of an environment for the driver while controlling downstream effects of high beam glare for active vehicles.

Systems and methods for compressing images that include a memory storing an executable code and a processor executing the code to receive a whole slide image, the whole slide image containing a plurality of image layers and metadata associated with each image layer, extract a high-resolution image layer and the corresponding metadata, wherein the high-resolution image layer includes a plurality of image tiles including informative tiles and noninformative tiles, where the informative tiles depict a region of interest of the specimen, analyze the image tiles of the extracted high-resolution image layer, determine a first tile is a noninformative tile, create an informative image layer by removing the first tile from the extracted high-resolution image layer, the informative image layer containing a plurality of informative tiles, compress the informative image layer into a single-layer whole slide image, and save the single-layer whole slide image in the memory.

A classification training method for training classifiers adapted to identify specific mutations associated with different cancer including identifying driver mutations. First cells from mutation cell lines derived from conditions having the number of driver mutations are acquired and 3D image feature data from the number of first cells is identified. 3D cell imaging data from the number of first cells and from other malignant cells is generated, where cell imaging data includes a number of first individual cell images. A second set of 3D cell imaging data is generated from a set of normal cells where the number of driver mutations are expected to occur, where the second set of cell imaging data includes second individual cell images. Supervised learning is conducted based on cell line status as ground truth to generate a classifier.

Disclosed herein are CPM support systems, as well as related apparatuses, methods, computing devices, and computer-readable media. For example, in some embodiments, a charged particle microscope computational support apparatus may include: first logic to, for each angle of a plurality of angles, receive an associated image of a specimen at the angle, and generate an associated scan mask based on one or more regions-of-interest in the associated image; second logic to, for each angle of the plurality of angles, generate an associated data set of the specimen by processing data from a scan, in accordance with the associated scan mask, by a charged particle microscope of the specimen at the angle; and third logic to provide, for each angle of the plurality of angles, the associated data set of the specimen to reconstruction logic to generate a three-dimensional reconstruction of the specimen.

Provided are a leukocyte detection method, a system, an electronic device and a computer readable medium. The method comprises: acquiring a microcirculation image (S1); determining a location of an intra-tubular space of a capillary vessel from the microcirculation image (S2); and determining a leukocyte index based on image information of the intra-tubular space of the capillary vessel (S3).

A method of determining a coverage of an image by an apparatus including processing circuitry includes executing, by the processing circuitry, instructions that cause the apparatus to generate a first segmentation mask by segmenting an image, generate a modified mask by applying a morphological operation to the first segmentation mask, generate a modified masked input based on the image and an inversion of the modified mask, generate a second segmentation mask by segmenting the modified masked input, and determine a coverage of the image based on the first segmentation mask and the second segmentation mask.

Various embodiments of the present disclosure are directed towards a method for generating a risk group classification for an African American (AA) patient. The method includes extracting a first plurality of architectural features from a digitized H&E slide image of the AA patient. A risk score for the AA patient is generated based on the first plurality of architectural features, where the risk score is prognostic of overall survival (OS) of the AA patient. The risk group classification is generated for the AA patient, where generating the risk group classification includes classifying the AA patient into either a high risk group or a low risk group based on the risk score, where the high risk group indicates the AA patient will die before a threshold date and the low risk group indicates the AA patient will die after or on the threshold date.