Abstract: An image processing system includes a housing that forms a passage through which paper pieces pass, imaging units that capture images of the paper pieces passing through the housing, and an AI image processing apparatus that restores a restoration target on the basis of the images captured by the imaging unit so that the contents of the restoration target become viewable.

Abstract: A learning-based active surface model for medical image segmentation uses a method including: (a) data generation: obtaining medical images and associated ground truths, and splitting the sample images into a training set and a testing set; (b) raw segmentation: constructing a surface initialization network, parameters of the network trained by images and labels in the training set; (c) surface initialization: segmenting the images by the surface initialization network, and generating the point cloud data as the initial surface from the segmentation; (d) fine segmentation: constructing the surface evolution network, the parameters of the network trained by the initial surface obtained in step (c); (e) surface evolution: deforming the initial surface points along the offsets to obtain the predicted surface, the offsets presenting the prediction of the surface evolution network; (f) surface reconstruction: reconstructing the 3D volumes from the set of predicted surface points set to obtain the final segmentatio

Abstract: Various methods and systems are provided for a medical imaging system. In one embodiment, a method for a projection imaging system includes acquiring a first image of a region of interest (ROI) with the projection imaging system in a first position, determining a three-dimensional (3D) location of an annotation on the first image via a geometric transformation using planes, acquiring a second image of the ROI with the projection imaging system in a second position, determining a location of the annotation on the second image based on the 3D location of the annotation in the first position and a geometry of the second position, and displaying the annotation on the second image in response to an accuracy check being satisfied.

Abstract: A CADx system for analysing medical images and determining if the images are acceptable for analysis, by determining if the images contain out-of-distribution input data is described. The CADx system comprises: an input circuit for receiving at least one medical image; a gatekeeper circuit for determining if the at least one received medical image contains out-of-distribution input data and so does not meet the requirements of the CADx system for acceptable images; and an output circuit to produce an output that is either a determination that the at least one received medical image contains out-of-distribution data and is not suitable for analysis by the CADx system, or a determination that the medical image is acceptable.

Abstract: A method and system of generating agent and actions prediction based on multi-agent tracking data are disclosed herein. A computing system retrieves tracking data from a data store. The computing system generates a trained neural network by generating a plurality of training data sets based on the tracking data by converting each frame of data into a matrix representation of the data contained in the frame and learning, by the neural network, a start frame and end frame of each action contained in the frame and its associated actor. The computing system receives target tracking data associated with an event. The target tracking data includes a plurality of actors and a plurality of actions. The computing system generates, via the trained neural network, a target start frame and a target end frame of each action identified in the tracking data and a corresponding actor.

Abstract: A passage propriety assessment device, provided with: a assessment unit that authorizes a person to pass through a first entrance/exit when a first facial image of the person at the first entrance/exit satisfies an individual authorization criterion, which is a criterion for assessing by facial recognition that the person is a pass-authorized person; and a setting unit that, when it is assessed that a person who has been authorized to pass through the first entrance/exit is impersonating a pass-authorized person, sets the individual authorization criterion for a pass-authorized person for a second entrance/exit located after the first entrance/exit so as to be higher than the individual authorization criterion for the first entrance/exit. When a second facial image of a person at the second entrance/exit does not satisfy the individual authorization criterion that was set higher, the assessment unit restricts the person from passing through the second entrance/exit.

Abstract: A computer-implemented method of machine learning including learning a Convolutional Neural Network (CNN) architecture for estimating a degradation generated by a denoiser on a ray traced image. The method includes obtaining a dataset and learning the CNN architecture based on the obtained dataset. The learning including taking as input an image generated by the denoiser and a corresponding noisy image of the provided dataset and outputting an error map. This forms an improved solution with respect to estimating a degradation generated by a denoiser on a ray traced image.

Abstract: A traffic monitoring system includes a storage device that stores a program and a hardware processor, in which the hardware processor executes a program stored in the storage device, thereby acquiring a position of a mobile object based on information from a detection device for detecting the position of the mobile object, generating information on a density distribution in which index values having a distribution according to the position of the mobile object are superimposed on one another for a plurality of mobile objects, and predicting whether a target mobile object is likely to enter a gap between two of the mobile objects based on a temporal change in density value indicated by the information on a density distribution.

Abstract: Embodiments relate to performing inference, such as object recognition, based on sensory inputs received from sensors and location information associated with the sensory inputs. The sensory inputs describe one or more features of the objects. The location information describes known or potential locations of the sensors generating the sensory inputs. An inference system learns representations of objects by characterizing a plurality of feature-location representations of the objects, and then performs inference by identifying or updating candidate objects consistent with feature-location representations observed from the sensory input data and location information. In one instance, the inference system learns representations of objects for each sensor. The set of candidate objects for each sensor is updated to those consistent with candidate objects for other sensors, as well as the observed feature-location representations for the sensor.

Abstract: One or more multi-layer systems are used to perform inference. A multi-layer system may correspond to a node that receives a set of sensory input data for hierarchical processing, and may be grouped to perform processing for sensory input data. Inference systems at lower layers of a multi-layer system pass representation of objects to inference systems at higher layers. Each inference system can perform inference and form their own versions of representations of objects, regardless of the level and layer of the inference systems. The set of candidate objects for each inference system is updated to those consistent with feature-location representations for the sensors as well as object representations at lower layers. The set of candidate objects is also updated to those consistent with candidate objects from other inference systems, such as inference systems at other layers of the hierarchy or inference systems included in other multi-layer systems.

Abstract: A method and apparatus with spoofing consideration is provided. The method includes implementing convolution block(s) of a machine learning model that determines whether biometric information in an input image is spoofed, including generating a feature map including channels for an input feature map for the input image using convolution layers of a convolution block of the convolution block(s), in response to a total number of input channels of the convolution block and a total number of output channels of the convolution block being different, matching the total number of input channels of the convolution block and the total number of output channels of the convolution block by adding a zero-padding channel to the input feature map using a skip connection structure, and generating output data for determining whether the biometric information is spoofed, dependent on the generated feature map and a result of the skip connection structure.

Abstract: Disclosed are a profiling method and apparatus based on a personal region of interest. The profiling method includes determining a region of interest (ROI) indicative of an interest region of a user based on the profile of a person, and generating relationship information indicative of a relationship between a specific person and the user based on a degree of intimacy between persons and the ROI. The AI device and the AI system of the present disclosure may be associated with an artificial intelligence module, a drone (or unmanned aerial vehicle (UAV)), a robot, an augmented reality (AR) device, a virtual reality (VR) device, a device related to 5G service, etc.

Abstract: A super resolution is produced using multiple reference images. Reference images are upsampled and blurred as needed for comparison between images of different resolution. Patches in blurred images are searched to find those patches which can be assembled into vectors for improving feature content over multiple resolution levels. The searches are based on similarity maps. The assembled vectors are concatenated with one or more other vectors, up-converted and then passed through convolutional layers to obtain new feature vectors. A final feature vector is passed through a convolutional layer to obtain the super resolution image.

Abstract: A method of generating a virtual 3D model of a dental arch is provided. The method includes receiving intraoral scans of a dental arch, determining a first depth of a first intraoral 3D surface in a first intraoral scan, and determining a second depth of a second intraoral 3D surface in the first intraoral scan, and wherein there is a fixed distance between the first intraoral 3D surface and the second intraoral 3D surface in the first intraoral scan. The method further includes stitching together the intraoral scans and generating a virtual 3D model of the dental arch from the intraoral scans, wherein the fixed distance between the first intraoral 3D surface and the second intraoral 3D surface is included in the virtual 3D model.

Abstract: Existing, low quality images can be restored using reconstruction or a combination of post-reconstruction techniques to generate a real patient phantom. The real patient phantom (RPP) can then be simulated in Monte Carlo simulations of a higher performance system and a lower performance system. Alternatively, the RPP can be simulated in the higher performance system, and a real scan can be performed by an existing, lower performance system. The higher performance system can be differentiated from the lower performance system in a variety of ways, including a higher resolution time of flight measurement capability, a greater sensitivity, smaller detector crystals, or less scattering. A neural network can be trained using the images produce by the higher performance system as the target, and the images produced by the lower performance system as the input.

Abstract: This disclosure describes one or more implementations of a video inference system that utilizes machine-learning models to efficiently and flexibly process digital videos utilizing various improved video inference architectures. For example, the video inference system provides a framework for improving digital video processing by increasing the efficiency of both central processing units (CPUs) and graphics processing units (GPUs). In one example, the video inference system utilizes a first video inference architecture to reduce the number of computing resources needed to inference digital videos by analyzing multiple digital videos utilizing sets of CPU/GPU containers along with parallel pipeline processing. In a further example, the video inference system utilizes a second video inference architecture that facilitates multiple CPUs to preprocess multiple digital videos in parallel as well as a GPU to continuously, sequentially, and efficiently inference each of the digital videos.

Abstract: Systems, methods, and other embodiments described herein relate to evaluating a perception network in relation to the accuracy of depth estimates and object detections. In one embodiment, a method includes segmenting range data associated with an image according to bounding boxes of objects identified in the image to produce masked data. The method includes comparing the masked data with corresponding depth estimates in the depth map according to an evaluation mask that correlates the depth estimates with the depth map. The method includes providing a metric that quantifies the comparing to assess a network that generated the depth map and the bounding boxes.

Abstract: An Enterobius vermicularis detection system is provided comprising a sample substrate, a sample disposed upon the sample substrate, and a personal computing device comprising an imaging device for acquiring one or more digital images of the sample substrate. The personal computing device may operate a digital interface allowing the acquired one or more digital images to be uploaded to one or more analysis computation servers. The one or more analysis computation servers may execute an analysis process upon the uploaded one or more digital images allowing determination of a presence of one or more Enterobius vermicularis eggs within the sample. The analysis process may output a first confidence interval value representing a likelihood of the presence of one or more Enterobius vermicularis eggs within the sample. The one or more analysis computation servers may communicate the first confidence interval value to the personal electronic device.

Abstract: A computer implemented network for executing a self-supervised scene change detection method in which image pairs (T0, T1) from different time instances are subjected to random photometric transformations to obtain two pairs of augmented images (T0?T0?, T0?; T1?T1?, T1?), which augmented images are passed into an encoder (f?) and a projection head (g?) to provide corresponding feature representations.

Abstract: A method and a system for providing a search result for similar products based on deep learning receive a search image including a product to be searched; search similar product images by performing deep learning based on the search image; when a plurality of the similar product images are searched, performing a first arrangement based on a target similarity for each of the plurality of similar product images; performing deep learning on the plurality of the similar product images, arranged based on the target similarity, based on one or more predetermined parameters; obtaining a mutual similarity representing a similarity between at least two of the plurality of similar product images by performing the deep learning based on the predetermined parameter; obtaining an integrated similarity for each of the plurality of similar product images based on the mutual similarity and the target similarity; performing a second arrangement by arranging the plurality of similar product images based on the integrated simil