Abstract: An object management system includes an identifier generation device and an object collation device. The identifier generation device includes a generation unit that forms an ink layer on a target object, an imaging unit that images an uneven pattern on a surface of the ink layer, and a registration unit that registers the imaged result in a storage unit. The object collation device includes an imaging unit that images the uneven pattern on the surface of the ink layer formed on the target object, and a recognizing unit that recognizes the target object based on an image of the uneven pattern obtained by imaging.

Abstract: A method for colourising a three-dimensional point cloud including surveying a point cloud with a surveying instrument. Each point of the point cloud may be characterised by coordinates within an instrument coordinate system having an instrument center. The method may include capturing a first image of the setting with a first camera. Each pixel value of the first image is assigned coordinates within a first camera coordinate system having a first projection center as origin and a first parallax shift relative to the instrument center. The method may include transforming the point cloud from the instrument coordinate system into the first camera coordinate system, resulting in a first transformed point cloud, detecting one or more uncovered points within the first transformed point cloud which are openly visible from the first projection center, and for each uncovered point, assigning a pixel value having corresponding coordinates in the first camera coordinate system.

Abstract: The present invention discloses an automatic pancreas CT segmentation method based on a saliency-aware densely connected dilated convolutional neural network. Under a coarse-to-fine two-step segmentation framework, the method uses a densely connected dilated convolutional neural network as a basis network architecture to obtain multi-scale image feature expression of the target. An initial segmentation probability map of the pancreas is predicted in the coarse segmentation stage. A saliency map is then calculated through saliency transformation based on a geodesic distance transformation. A saliency-aware module is introduced into the feature extraction layer of the densely connected dilated convolutional neural network, and the saliency-aware densely connected dilated convolutional neural network is constructed as the fine segmentation network model. A coarse segmentation model and the fine segmentation model are trained using a training set, respectively.

Abstract: A method (900) of generating an atlas for a universal atlas database (901) is provided. A new medical scan image (905) is provided. A universal auto-contouring operation (920) is performed on the medical scan image, to generate a set of universal contours (930) for the medical scan image. A local auto-contouring customisation operation (940) is performed on the medical scan image, to generate a set of local contours (950) for the medical scan image. The set of local contours is standardised (980) using a trained model to compensate for biases in the set of local contours, thereby creating a set of standardised global contours (985) for the medical scan image. The set of standardised global contours (985) and the medical scan image (905) can be added to the universal atlas database (901) as a new atlas, thereby expanding the set of atlases that are available in the universal atlas database.

Abstract: The image-capturing unit includes an imaging section; a holding section configured to hold a subject to be imaged by the imaging section; a light irradiation section configured to select light of one or more light sources out of multiple light sources having different wavelengths, and to irradiate the subject held in the holding section with the light; a storage section configured to store a correspondence among a color of the light emitted for irradiating the subject by the light irradiation section, a material of an irradiation surface irradiated with the light, and a resolution representing the number of pixels per unit length; and an image processing section configured to obtain the resolution from the correspondence, based on the color of the light emitted for irradiating the subject and the material of the irradiation surface of the subject, and to process a subject image by using the resolution.

Abstract: A method for analyzing biological-tissue image includes following steps. A plurality of biological-tissue images are provided, and each of the biological-tissue images includes a plurality of target object image blocks. An image pre-processing step is performed so as to obtain a plurality of processed biological-tissue images. A fitting step is performed so as to obtain a plurality of object fitting images of the target object image blocks. A sampling step is performed, wherein a target region of each of the processed biological-tissue images is selected, and the target region includes the object fitting images. A calculating and analyzing step is performed so as to obtain an analysis result of a target regional center of the biological-tissue images.

Abstract: Visual target tracking is task of locating a target in consecutive frame of a video. Conventional systems observe target behavior frames of the video. However, dealing with this problem is very challenging when video has illumination variations, occlusion, change in size and view of the object due to relative motion between camera and object. Embodiments of the present disclosure addresses this problem by implementing Neural Network (NN), its features and their corresponding gradients. Present disclosure explicitly guides the NN by feeding target object of interest (ToI) defined by a bounding box in the first frame of the video. With this guidance, NN generates target activation map via convolutional features map and their gradient maps, thus giving tentative location of the ToI to further exploit to locate target object precisely by using correlation filter(s) and peak location estimator, thus repeating process for every frame of video to track ToI accurately.

Abstract: An endoscope system includes a processor including an image processing unit that obtains an inflammation evaluation value in which a degree of inflammation of a biological tissue is digitized on the basis of information of a color component of an image of the biological tissue, from the image that is obtained by imaging the biological tissue, and a monitor displaying the inflammation evaluation value. The image processing unit includes a blood vessel region determination unit obtaining certainty of a blood vessel region of the biological tissue in the image, a pixel evaluation value generation unit obtaining a pixel evaluation value by performing digitization processing with respect to each of the pixels of the image, a pixel evaluation value adjustment unit calculating an adjustment value in which the pixel evaluation value is reduced as the certainty of the blood vessel region increases.

Abstract: An image processing apparatus includes a processor including hardware. The processor is configured to: calculate a motion evaluation value on a motion of a subject in a determination target frame; acquire a difference evaluation value of the determination target frame; determine whether to perform defective pixel detection on an image of the determination target frame by using the motion evaluation value and the difference evaluation value; and when it is determined that the defective pixel detection is to be performed, detect a defective pixel by determining whether a pixel of interest that is a determination target is a defective pixel with respect to the image of the determination target frame, based on a pixel value of the pixel of interest and pixel values of neighboring pixels that are located in a vicinity of the pixel of interest.

Abstract: A system and a method for monitoring a brain CT scan image using ASPECTS score. The method includes receiving the brain CT scan image of a patient. Further, a basal ganglia region and a corona radiata level are identified in a plurality of slices in the brain CT scan image. Furthermore, a plurality of anatomical regions and a plurality of infarcts are segmented using deep learning. Subsequently, an overlapping region across the plurality of slices is determined based on the plurality of anatomical regions and the plurality of infarcts. The overlapping region and a predefined threshold are used to compute an ASPECTS score. The ASPECTS score is further used to recommend a course of action to the patient.

Abstract: Systems and methods are disclosed for receiving a digital image corresponding to a target specimen associated with a pathology category, wherein the digital image is an image of tissue specimen, determining a detection machine learning model, the detection machine learning model being generated by processing a plurality of training images to output a cancer qualification and further a cancer quantification if the cancer qualification is an confirmed cancer qualification, providing the digital image as an input to the detection machine learning model, receiving one of a pathological complete response (pCR) cancer qualification or a confirmed cancer quantification as an output from the detection machine learning model, and outputting the pCR cancer qualification or the confirmed cancer quantification.

Abstract: A method of embedding a bit or more of watermark data in a video signal to be entropy coded, wherein the video signal may be a block with levels, wherein the method may include: obtaining the watermark data; and obtaining a value of a first watermarked level on the basis of a first level of the block by processing the value of the first watermarked level. There also is a method of detecting a watermark in a video signal, the video signal having been entropy decoded and with a block with a first watermarked level, wherein the method may include: receiving the video signal; obtaining at least one bit of watermark data wi? on the basis of the first watermarked level; and determining a tampering indicator by verifying whether the obtained watermark data wi? corresponds with watermark data wi.

Abstract: A gauze detection system is provided that is capable of effectively detecting a gauze pad in the patient's body during surgery without applying special processing to the gauze pad. A gauze detection system 100, includes: an image input section 1 to input a taken image of an operative field; a determination section 2 to determine whether a determination target region contains a feature of a gauze image by image processing, the region having a predetermined size in an input image; and a determination result output section 3 to report detection of a gauze pad in the input image in a case of the determination target region being determined by the determination section 2 to contain the feature of the gauze image.

Abstract: Provided is a dynamic analysis apparatus that predicts a respiratory function value based on frame images showing dynamics of chest. The dynamic analysis apparatus includes a hardware processor that obtains a first lung size value of a removal target site and a second lung size value of a left or right lung field including the removal target site, calculates a proportion between the first and second lung size values as a size proportion, calculates a first feature amount concerning respiratory function of the left or right lung field including the removal target site and a second feature amount concerning respiratory function of the lung fields as a whole, calculates a proportion between the first and second feature amounts as a feature amount proportion, and calculates an expected rate of the respiratory function without the removal target site, based on a product of the size proportion and the feature amount proportion.

Abstract: Deep learning approaches automatically segment at least some breast tissue images while non-deep learning approaches automatically segment organs-at-risk. Both three-dimensional CT imaging information and two-dimensional orthogonal topogram imaging information can be used to determine virtual-skin volume. The foregoing imaging information can also serve to automatically determine a body outline for at least a portion of the patient. That body outline, along with the virtual-skin volume and registration information can serve as inputs to automatically calculate radiation treatment platform trajectories, collision detection information, and virtual dry run information of treatment delivery per the optimized radiation treatment plan.

Abstract: A method includes calculating a medial-axis tree graph of a volume of an organ of a patient in a computerized anatomical map of the volume. A predefined number of major branches in the tree graph are identified. Using the identified major branches, one or more known anatomical opening regions of the volume are identified in the anatomical map.

Abstract: Systems and methods are disclosed for adjusting attributes of whole slide images, including stains therein. A portion of a whole slide image comprised of a plurality of pixels in a first color space and including one or more stains may be received as input. Based on an identified stain type of the stain(s), a machine-learned transformation associated with the stain type may be retrieved and applied to convert an identified subset of the pixels from the first to a second color space specific to the identified stain type. One or more attributes of the stain(s) may be adjusted in the second color space to generate a stain-adjusted subset of pixels, which are then converted back to the first color space using an inverse of the machine-learned transformation. A stain-adjusted portion of the whole slide image including at least the stain-adjusted subset of pixels may be provided as output.

Abstract: Systems and methods for reconstructing medical images are disclosed. Measurement data, such as sinogram data, is received from an image scanning system. A plurality of masks are applied to corresponding portions of the measurement data to generate a plurality of masked measurement data portions. In some examples, the measurement data is encoded before the plurality of masks are applied. A neural network including a plurality of fully connected layers is applied to the plurality of masked measurement data portions to generate a plurality of image patches. The plurality of image patches are then combined to generate an initial image. In some examples, refinement and scaling operations are applied to the initial image and corresponding attenuation maps to generate a final image. In some examples, the final image is stored in a database. In some examples, the final image is displayed for diagnosis.

Abstract: Systems and methods are disclosed for adjusting attributes of whole slide images, including stains therein. A portion of a whole slide image comprised of a plurality of pixels in a first color space and including one or more stains may be received as input. Based on an identified stain type of the stain(s), a machine-learned transformation associated with the stain type may be retrieved and applied to convert an identified subset of the pixels from the first to a second color space specific to the identified stain type. One or more attributes of the stain(s) may be adjusted in the second color space to generate a stain-adjusted subset of pixels, which are then converted back to the first color space using an inverse of the machine-learned transformation. A stain-adjusted portion of the whole slide image including at least the stain-adjusted subset of pixels may be provided as output.

Abstract: Super-resolution images are generated from standard-resolution images acquired with a magnetic resonance imaging (“MRI”) system. More particularly, super-resolution (e.g., sub-millimeter isotropic resolution) images are generated from standard-resolution images (e.g., images with 1 mm or coarser isotropic resolution) using a deep learning algorithm, from which accurate cortical surface reconstructions can be generated.