Abstract: A method (and system) of segmenting one or more histological structures in a tissue image represented by multi-parameter cellular and sub-cellular imaging data includes receiving coarsest level image data for the tissue image, wherein the coarsest level image data corresponds to a coarsest level of a multiscale representation of first data corresponding to the multi-parameter cellular and sub-cellular imaging data. The method further includes breaking the coarsest level image data into a plurality of non-overlapping superpixels, assigning each superpixel a probability of belonging to the one or more histological structures using a number of pre-trained machine learning algorithms to create a probability map, extracting an estimate of a boundary for the: one or more histological structures by applying a contour algorithm to the probability map, and using the estimate of the boundary to generate a refined boundary for the one or more histological structures.

Abstract: A method for generating a module configured to determine concentration of an analyte in a sample of a body fluid is disclosed. The method includes providing a first set of measurement data derived from images of one or more test strips indicating a color transformation in response to a body fluid containing an analyte. The images can be recorded by multiple devices with differing cameras, software and/or hardware device configurations for image recording and image data processing. A neural network model can be generated in a machine learning process applying an artificial neural network and a module configured to determine concentration of an analyte in a second sample of a body fluid can be generated. Further, the present disclosure includes a system for generating the module as well as a method and a system for determining concentration of an analyte in a sample of a bodily fluid.

Abstract: The present development is a method for providing stable visualization of tubular objects. The method simplifies visualization using a Fly-Ln approach without loss of internal surface details. The method uses graphical programming, and can be easily integrated with commercially available visualization programs, such as Virtual Colonoscopy (VC) or Computed Tomography Colonography (CTC).

Abstract: The invention refers to an apparatus for monitoring a subject (121) during an imaging procedure, e.g. CT-imaging. The apparatus (110) comprises a monitoring image providing unit (111) providing a first monitoring image and a second monitoring image acquired at different support positions, a monitoring position providing unit (112) providing a first monitoring position of a region of interest in the first monitoring image, a support position providing unit (113) providing support position data of the support positions, a position map providing unit (114) providing a position map mapping calibration support positions to calibration monitoring positions, and a region of interest position determination unit (115) determining a position of the region of interest in the second monitoring image based on the first monitoring position, the support position data, and the position map. This allows to determine the position of the region of interest accurately and with low computational effort.

Abstract: This machine-learning device is provided with: a detection unit which detects a loss of consistency with a lapse of time in a determination result for unit data, the determination result being output from a determination unit that generates a learning model to be used when performing prescribed determination for one or more pieces of the unit data that form time series data; and a selection unit which selects, on the basis of the result of detection by the detection unit, unit data to be used as teacher data when the determination unit updates the learning model, thereby efficiently raising the accuracy of the learning model when machine learning is performed on the basis of the time series data.

Abstract: A care system and an automatic care method are provided. The care system is suitable for a care recipient. A physiological condition of the care recipient can be determined by sensors arranged around the care recipient, and a health condition of the care recipient can be determined through monitoring sleep disorders thereof. The care system receives a sound of the care recipient through a sound receiver, after removing a background sound, a sound of the care recipient can be obtained, and an image sensor is used to obtain an image of the care recipient to perform a motion detection, so as to obtain a posture image of the care recipient. Artificial intelligence technology can be used to build a care-taking warning prediction model to determine whether or not the status of the care recipient has reached a warning threshold.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network to perform a downstream computer vision task. One of the methods includes pre-training an initial neural network that shares layers with the neural network to perform an initial computer vision task and then training the neural network on the downstream computer vision task.

Abstract: Systems and methods for augmenting a training data set with annotated pseudo images for training machine learning models. The pseudo images are generated from corresponding images of the training data set and provide a realistic model of the interaction of image generating signals with the patient, while also providing a realistic patient model. The pseudo images are of a target imaging modality, which is different than the imaging modality of the training data set, and are generated using algorithms that account for artifacts of the target imaging modality. The pseudo images may include therein the contours and/or features of the anatomical structures contained in corresponding medical images of the training data set. The trained models can be used to generate contours in medical images of a patient of the target imaging modality or to predict an anatomical condition that may be indicative of a disease.

Abstract: A system for visualizing a surgical site is provided. The system includes a robotic mechanism for performing a procedure on a patient, an imaging device coupled to the robotic mechanism, the imaging device configured to provide image data of a site of interest, and a computing device coupled to the imaging device. The computing device includes one or more processors and at least one memory device configured to store executable instructions. The executable instructions, when executed by the processor, are configured to receive the image data of the site of interest, track motion patterns of the site of interest in the received image data, filter the received image data to remove line-of-sight restrictions therein and alter pixels therein based on the tracked motion patterns, and generate an output frame from the filtered image data. The system also includes a presentation interface device coupled to the computing device and configured to present the output frame for visualization of the site of interest.

Abstract: A system is provided that includes a Q3D endoscope disposed to image a field of view and a processor that produces a Q3D model of a scene and identifies target instruments and structures. The processor is configured to display the scene from a virtual field of view of an instrument, to determine a no fly zone around targets, to determine a predicted path for said instruments or to provide 3D tracking of said instruments.

Abstract: Systems and methods are disclosed for receiving a target electronic image corresponding to a target specimen, the target specimen comprising a tissue sample of a patient, applying a machine learning system to the target electronic image to identify a region of interest of the target specimen and determine an expression level of, category of, and/or presence of a biomarker in the region of interest, the biomarker comprising at least one from among an epithelial growth factor receptor (EGFR) biomarker and/or a DNA mismatch repair (MMR) deficiency biomarker, the machine learning system having been generated by processing a plurality of training images to predict whether a region of interest is present in the target electronic image, the training images comprising images of human tissue and/or images that are algorithmically generated, and outputting the determined expression level of, category of, and/or presence of the biomarker in the region of interest.

Abstract: A mobile radiographic imaging apparatus that performs dynamic imaging by using radiation, includes a detector, a wireless communication interface, and a hardware processor. The detector detects a first wireless access point and a second wireless access point. The wireless communication interface connects to the first wireless access point as a connection point and outputs a dynamic image including a plurality of frames acquired in the dynamic imaging to the first wireless access point. The hardware processor controls a change of the connection point from the first wireless access point to the second wireless access point. The hardware processor further performs control to inhibit the change of the connection point in a period when the frame images are still being output to the first wireless access point.

Abstract: In a case where a first specific scene recognized at a time of insertion of an endoscope is stored in a specific scene memory and a recognized scene recognized by a scene recognition unit is a scene on a side deeper than the first specific scene in a direction of movement of the endoscope, the recognized scene is output without being changed. In a case where the first specific scene is stored in the specific scene memory and the recognized scene is a scene on a side shallower than the first specific scene in the direction of movement of the endoscope, the recognized scene is changed and output.

Abstract: Described herein are Deep Multi-Magnification Networks (DMMNs). The multi-class tissue segmentation architecture processes a set of patches from multiple magnifications to make more accurate predictions. For the supervised training, partial annotations may be used to reduce the burden of annotators. The segmentation architecture with multi-encoder, multi-decoder, and multi-concatenation outperforms other segmentation architectures on breast datasets, and can be used to facilitate pathologists' assessments of breast cancer in margin specimens.

Abstract: Systems and methods provide automated systems for accurately estimating infusion coverage within a target brain region in real-time (or approximately real-time) during surgical procedures. Such accurate and real-time infusion coverage estimation enables intraoperative monitoring and adjustment of infusion parameters (e.g., cannula tip location, infusate delivery flow rate, etc.) for achieving optimal/improved infusion coverage for a given drug therapy. Accordingly, examples of the presently disclosed technology can improve the efficacy and safety of drug therapies delivered to the brain.

Abstract: A computer-implemented method of estimating a pose of a target object in a three-dimensional scene includes: obtaining image data and associated depth information representing a view of the three-dimensional scene; processing the image data and the associated depth information to generate a volumetric reconstruction for each of a plurality of objects in the three-dimensional scene, including the target object; determining a volumetric grid containing the target object; generating, using the generated volumetric reconstructions, occupancy data indicating portions of the volumetric grid occupied by free space and portions of the volumetric grid occupied by objects other than the target object; and estimating the pose of the target object using the generated occupancy data and pointwise feature data for a plurality of points on a surface of the target object.

Abstract: A computer-implemented method of reducing scatter in an X-ray projection image of an object comprises: generating an initial X-ray projection image with an imaging beam and an X-ray detector; based on a first position in a detector array of the X-ray detector, selecting a first kernel for convolution of a first portion of the initial projection image, wherein the first position corresponds to the first portion of the initial projection image; based on a second position in the detector array of the X-ray detector, selecting a second kernel for convolution of a second portion of the initial projection image, wherein the second position corresponds to the second portion of the initial projection image; convolving the first portion with the first kernel and the second portion with the second kernel to generate a scatter component of the initial X-ray projection image; and generating a corrected X-ray projection image by removing the scatter component from the initial X-ray projection image.

Abstract: Systems and methods are provided for augmenting digital analysis of lesions. An image of tissue having a glandular epithelial component is generated. The image represents a plurality of medium-scale epithelial components. For each of a plurality of cells within the image, a representative point is identified to provide a plurality of representative points for each of the plurality of medium-scale epithelial components. For each of a subset of the plurality of medium-scale epithelial components, a graph connecting the plurality of representative points is constructed. A plurality of classification features is extracted for each of the subset of medium-scale epithelial components from the graph constructed for the medium-scale epithelial component. A clinical parameter is assigned to each medium-scale epithelial component according to the extracted plurality of classification features.

Abstract: A method for modeling a joint before, during, and/or after a medical procedure includes receiving first imaging data capturing the joint from a first imaging perspective and second imaging data capturing the joint from a second imaging perspective that is different than the first imaging perspective, the first and second imaging data generated intraoperatively via a two-dimensional imaging modality, generating three-dimensional image data by back-projecting the first and second imaging data in three-dimensional space in accordance with a relative difference between the first and second imaging perspectives, generating a three-dimensional model of the joint based on processing the three-dimensional image data with a machine learning model trained on imaging data generated via at least a three-dimensional imaging modality, and displaying a visualization based on the three-dimensional model of the joint during the medical procedure.

Abstract: Described herein are means for generating source models for transfer learning to application specific models used in the processing of medical imaging.

Abstract: The present application discloses a method, device, and system for processing a medical image. The method includes obtaining, by one or more processors, a first medical image, wherein the first medical image includes a representation of a target organ, and processing, by the one or more processors, the first medical image based at least in part on a first machine learning model. A classification result associated with the first medical image and a first image of a target region in the first medical image are obtained based at least in part on the processing of the first medical image using the first machine learning model.

Abstract: A shade matching method for a tooth obtains a plurality of digitized tooth shade references, obtains a three-dimensional (3D) surface representation of a tooth and associated color information defined on the three-dimensional (3D) surface, selects a region on the three-dimensional (3D) surface representation of the tooth, determines a plurality of correspondences with the common tooth shape of all the tooth shade references, computes color difference with each of the tooth shade reference from the plurality of tooth shade references, and records the label corresponding to the shade reference with a smallest difference.

Abstract: A method comprises obtaining, for one or more medical instruments coupled to a teleoperational medical system, one or more corresponding medical instrument positions relative to a field of view of at least one imaging instrument coupled to the teleoperation medical system. The method also comprises initiating, for a selected medical instrument of the one or more medical instruments, generation of a corresponding indicator to a user. The corresponding indicator may be based on a position of the selected medical instrument relative to the field of view.

Abstract: A personalized fixation system includes a surgical planning software tool configured to adjust relationships of relevant anatomy of a subject, at least one bone anchor, a plate having a shape that does not conform to a single plane, the plate configured to accept the at least one bone anchor, wherein the shape of the plate is at least partially determined by the surgical planning software tool, and wherein the plate includes at least one node having a hole configured to receive the at least one bone anchor, and a locking element configured to connect the at least one bone anchor to the plate, wherein the plate is manufactured using additive manufacturing.

Abstract: A system may comprise a tool including at least one reference feature. a processor, and a memory having computer readable instructions stored thereon. The computer readable instructions, when executed by the processor, may cause the system to receive image data including an image of the tool and the at least one reference feature, determine a pose of the tool from the image data, and modify the image data to visually decrement a portion of the image data corresponding to the at least one reference feature.

Abstract: A method for registering a two-dimensional image data set with a three-dimensional image data set of a body of interest is discloses herein. The method includes the following steps: defining a first vector of a registered virtual camera in a coordinate system of the three-dimensional image data et; obtaining a first transformed vector of an unregistered virtual camera in the coordinate system of the three-dimensional image data set by transforming the first vector of the registered virtual camera through at least one transforming matrix; defining a focal point of the unregistered virtual camera at a reference point of the two-dimensional image data set in the coordinate system of the three-dimensional image data set; and repositioning the unregistered virtual camera according to the first transformed vector and the focal point of the unregistered virtual camera.

Abstract: A radiological apparatus including: a gantry encapsulated within a cover, a patient platform, and two radiation sources with imaging directions orthogonal to each other, sliding vertically to perform vertical scanning of a patient standing on the platform. The gantry cover top view is L shaped, each radiation source being located outside the gantry cover, inside the angular sector of the L, and is encapsulated within a cover sliding vertically with the radiation source it encapsulates. The radiological apparatus also includes: a first security device stopping the vertical scanning, when it detects a patient body part going outside a first predetermined area, to avoid collision with the vertically sliding radiation sources covers, and a second security device stopping the vertical scanning, when it detects an object or a person external to the radiological apparatus within a second predetermined area, to avoid collision with the vertically sliding radiation sources covers.

Abstract: The disclosed calibration method includes a calibration phantom positioned on an adjustable table on the surface of a mechanical couch, with the phantom's centre at an estimated location for the iso-centre of a radio therapy treatment apparatus. The calibration phantom is then irradiated using the apparatus, and the relative location of the center of the calibration phantom and the iso-centre of the apparatus is determined by analyzing images of the irradiation of the calibration phantom. The calibration phantom is then repositioned by the mechanical couch applying an offset corresponding to the determined relative location of the centre of the calibration phantom and the iso-centre of the apparatus to the calibration phantom. Images of the relocated calibration phantom are obtained, to which the offset has been applied, and the obtained images are processed to set the co-ordinate system of a stereoscopic camera system relative to the iso-centre of the apparatus.

Abstract: A method (700) for localizing inflammation within a user's mouth using an oral care device (10), comprising the steps of: (i) emitting (720) light by one or more light emitters (42) of the oral care device; (ii) detecting (730), by a light detector (40) of the oral care device, reflectance from a surface at each of a plurality of locations within the user's mouth to generate reflectance data for each of the plurality of locations; (iii) determining (750), by the controller, for the plurality of locations, whether gingiva at that location is inflamed; and (iv) providing (760), via a user interface (46) of the oral care device, information to the user regarding the inflamed location or locations.

Abstract: To provide a medicine collation device, a medicine collation system, a medicine management method, and a program, capable of collecting medicine images that are candidates for master images. The medicine collation device includes: a medicine image acquiring unit configured to acquire a medicine image generated by imaging a medicine to be collated; a first master image storing unit configured to store a master image of the medicine; a collating unit configured to collate the medicine image with the master image; an associating unit configured to associate the medicine image collated by the collating unit, with identification information on the medicine image; and a transmitting unit configured to transmit the medicine image and the identification information that are associated in the associating unit, to an outside.

Abstract: In a computer-implemented method of iteratively reconstructing images or a volume using deep learning and a computer-implemented method of training, projection filter parameter(s) of a projection (joint) bilateral filter, (J)BF, are automatically tuned by a first trained neural network based on data of projections of an imaging procedure. Volume filter parameter(s) of a volume (J)BF are automatically tuned by a second trained neural network based on data of a volume backward-projected from projections filtered by the projection (joint) bilateral filter using the automatically tuned projection filter parameter(s). A volume filtered by the volume (J)BF is output as the reconstructed images or volume, in case the filtered volume meets a predefined quality criterion.

Abstract: One embodiment of a method for rendering one or more graphics images includes tracing one or more rays through a graphics scene; computing one or more surface normals associated with intersections of the one or more rays with one or more surfaces, where computing each surface normal includes: computing a plurality of intermediate surface normals associated with a plurality of adjacent voxels of a grid, and interpolating the plurality of intermediate surface normals; and rendering one or more graphics images based on the one or more surface normals.

Abstract: According to one embodiment, a reconstruction apparatus obtains acquired raw data, and reconstruct a data set that represents a measured physical amount with a multidimensional space based on the acquired raw data. Herein, the apparatus performs reduction processing for generating noise-reduced partial data from partial data relating to a partial area of a data set at a current number of iterations, error compensation processing for compensating errors in a data set at the current number of iterations with respect to the acquired raw data, and optimization processing for reconstructing the data set by repeating the reduction processing and the error compensation processing until predetermined criteria are met, while moving the partial area.

Abstract: The present application is suitable for use in the technical field of computers, and provides a smart diagnosis assistance method and terminal based on medical images, comprising: acquiring a medical image to be classified; pre-processing the medical image to be classified to obtain a pre-processed image; and inputting the pre-processed image into a trained classification model for classification processing to obtain a classification type corresponding to the pre-processed image, the classification model comprising tensorized network layers and a second-order pooling module. As the trained classification model comprises tensor decomposed network layers and a second-order pooling module, when processing images on the basis of the classification model, more discriminative features related to pathologies can be extracted, increasing the accuracy of medical image classification.

Abstract: Systems, methods, and devices for reconstructing an image is provided. An imaging device may be oriented at one or more poses and an image may be received at each of the one or more poses to form a set of images. Pose information may be received at each of the one or more poses. The set of images and the pose information may be inputted into a reconstruction model to generate a three-dimensional representation of the one or more anatomical elements.

Abstract: Devices, systems, and methods used to register medical image data with anatomical positions are disclosed. The image data is captured by an imaging system over a period of time. The positions are determined by a tracking system over the period of time. The positions are averaged to determine an average position of the anatomy, which is then used for registration of the image data with the anatomy. Registration based on the average position can then be used to enhance the quality of 3D images of the anatomy, as well as to improve navigation for a surgical procedure.

Abstract: A pharmaceutical composition for treating an influenza virus infectious disease, containing an anti-influenza virus agent as an active ingredient is provided. The anti-influenza virus agent is administered to a patient determined to be positive for the influenza virus infectious disease based on an intraoral image captured using an intraoral imaging apparatus.

Abstract: Embodiments are directed to use of a rapid-test-validation computing device to determine if a result of a rapid test device is valid and identify the result. The rapid-test-validation computing device captures images of the rapid test device and employs a first artificial intelligence mechanism to determine if the rapid test device is properly aligned in the images. The rapid-test-validation computing device then employs a second artificial intelligence mechanism to determine if a result of the rapid test device is valid or invalid. If the result is valid, the rapid-test-validation computing device employs a third artificial intelligence mechanism to determine and present an objective output of the rapid test device result to a user; otherwise, the rapid-test-validation computing device presents a notification to the user that the rapid test device result is invalid.

Abstract: A system and/or method uses a trained U-Net neural network to remove flow artifacts from optical coherence tomography (OCT) angiography (OCTA) data. The trained U-Net receives as input both OCT structural volume data and OCTA volume data, but expands the OCTA volume data to include depth information. The U-Net applies dynamic pooling along the depth direction and weighs more heavily the portion of the data that follows (e.g., along the contours of) select retinal layers. In this manner the U-Net applies contextually different computations at different axial locations based at least in part on the depth index information and/or (e.g., proximity to) the select retinal layers. The U-net outputs OCT volume data of reduced flow artifacts as compared with the input OCTA data.

Abstract: A method and system for measuring signal loss in a fiber optic cable. The tail ends of reference and test fiber optic cables are illuminated with a diffuse light. The head end of each of the reference and test fiber optic cables are positioned in a measurement area. A core imager captures an image of the core of each head-end while it is in the measurement area. Reference and test radiant fluxes emitted from the reference and test head-ends are determined from the respective core images. The relative signal loss of the test fiber optic cable is then determined by comparing the test radiant flux to the reference radiant flux.

Abstract: The present inventive concept relates to a non-face-to-face health status measurement system and method through a camera-based vital sign data extraction and an electronic questionnaire, providing a system and a method for monitoring various infectious diseases and enabling rapid response to the occurrence of diseases and infectious diseases by measuring a health status of a user in non-face-to-face through a vital sign data including heart rate, respiration rate, oxygen saturation, etc. extracted by using color information of a face image taken with a camera and results of an electronic questionnaire performed online.

Abstract: Techniques are described that facilitate dynamic multimodal segmentation selection and fusion in medical imaging. In one example embodiment, a computer processing system receives a segmentation dataset comprising a combination of different image segmentations of an anatomical object of interest respectively segmented via different segmentation models from different medical images captured of the (same) anatomical object, wherein the different medical images and the different image segmentations vary with respect to at least one of, capture modality, acquisition protocol, or acquisition parameters. The system employs a dynamic ranking protocol as opposed to a static ranking protocol to determine ranking scores for the different image segmentations that control relative contributions of the different image segmentations in association with combining the different image segmentations into a fused segmentation for the anatomical object.

Abstract: Techniques for segmentation include determining an edge of voxels in a range associated with a target object. A center voxel is determined. Target size is determined based on the center voxel. In some embodiments, edges near the center are suppressed, markers are determined based on the center, and an initial boundary is determined using a watershed transform. Some embodiments include determining multiple rays originating at the center in 3D, and determining adjacent rays for each. In some embodiments, a 2D field of amplitudes is determined on a first dimension for distance along a ray and a second dimension for successive rays in order. An initial boundary is determined based on a path of minimum cost to connect each ray. In some embodiments, active contouring is performed using a novel term to refine the initial boundary. In some embodiments, boundaries of part-solid target objects are refined using Markov models.

Abstract: An information processing apparatus including at least one processor, wherein the processor is configured to cause a learning model that is used to extract a region of interest from an input image, to perform learning in response to an input of a pseudo image generated in a case where a pixel value of at least a part of an original image obtained by imaging a subject is changed, as data for learning.

Abstract: A health management system using contactless physiological measurement technology is disclosed. The health management system principally comprises a camera and a first processor, of which the camera is faced to a user for capturing a user image. The first processor is particularly configured to have a face detection unit and an activity index calculating unit therein. By such arrangement, after receiving the user image from the camera, the first processor detects a face portion from the user image, thereby subsequently extracting a PPG signal from the face portion. Consequently, after completing at least one process of the PPG signal, multiple indexes for describing a user's health activity are generated. The health activity indexes include health index, activity index, stability index, relaxation index, metabolism index, and balance index. Therefore, the first processor achieves an evaluation of the user's health activity state according to the forgoing health activity indexes.

Abstract: Disclosed are a method and an apparatus of processing image. The method includes: obtaining an initial bone segmentation result; and fusing the initial bone segmentation result based on characteristics of and correspondences between a plurality of bone segmentation results in the initial bone segmentation result, to obtain a target bone segmentation result. The initial bone segmentation result includes the plurality of bone segmentation results generated by a plurality of different deep learning models. Methods in the embodiments of the present application can improve precision of a fusion result of the plurality of bone segmentation results.

Abstract: The present invention relates to a radiopharmaceutical distribution image generation system and method using deep learning and, more specifically, to a radiopharmaceutical distribution image generation system and method using deep learning, wherein dynamic medical images collected from multiple patients and a time-radiation dose distribution curve for each organ can be learnt through a deep learning network and a spatial distribution image of a radiopharmaceutical can be generated from a static medical image acquired from a specific patient. According to the present invention, even when a medical image is acquired only for a specific time after a radiopharmaceutical is injected into a patient, a spatial distribution image of the radiopharmaceutical can be acquired across the entire time by using the deep learning network, and quantitative analysis of the radiopharmaceutical can be performed by calculating a time-radiation dose distribution curve on the basis thereof.

Abstract: A system for screening persons' temperatures comprises an imaging device configured to generate imagery data, including infrared image data, of a first scene and a second scene. The system receives imagery data from the first scene, identifies persons in the scene using the imagery data, and determines each person's temperature. The system further compares the persons' temperatures to a threshold and indicates a person should be directed to the second scene if the person's temperature is above a threshold. Subsequently, the system determines and compares the person's temperature in the second scene and provides an indication if the person's temperature exceeds a second threshold. Further, a confidence factor can be associated with determinations of persons' temperatures.

Abstract: A computer-implemented method for determining an output signal characterizing a semantic segmentation and/or an instance segmentation of an image. The method includes: determining a first intermediate output signal from a machine learning system, wherein the first intermediate output signal characterizes a semantic segmentation and/or an instance segmentation of the image; adapting parameters of the machine learning system based on a loss function, wherein the loss function characterizes an entropy or a cross-entropy of the first intermediate output signal; determining the output signal from the machine learning system based on the image and the adapted parameters.

Abstract: Systems and methods described herein relate, among other things, to unmixing more than three stains, while preserving the biological constraints of the biomarkers. Unlimited numbers of markers may be unmixed from a limited-channel image, such as an RGB image, without adding any mathematical complicity to the model. Known co-localization information of different biomarkers within the same tissue section enables defining fixed upper bounds for the number of stains at one pixel. A group sparsity model may be leveraged to explicitly model the fractions of stain contributions from the co-localized biomarkers into one group to yield a least squares solution within the group. A sparse solution may be obtained among the groups to ensure that only a small number of groups with a total number of stains being less than the upper bound are activated.