Abstract: A method includes obtaining visual content comprising spatial portions; determining respective spatial qualities of the spatial portions, wherein the respective spatial qualities are based on locations of the spatial portions within the visual content; and encoding the spatial portions of the visual content based on the respective spatial qualities. An apparatus includes a camera, a display, and a processor. The processor is configured to identify, using facial recognition, a face of a user of the apparatus; identify a distance of the face of the user to the display; and render visual content on the display using a quality that is based on the distance.

Abstract: An augmented reality system enhances or augments the view of a real-world object by replacing different portions of the object with visualizations of non-visible data encoded for those different portions. The system may receive an image that captures visible properties of an object from a particular perspective. The system may select a model with data points that are generated from scanning the object and that encode visible and non-visible properties of the object. The system may align the model to match the object at the particular perspective, may map the non-visible properties from a set of the data points to an object part that is represented and aligned with the set of data points, and may present an augmented reality view in which the visible properties at the object part from the particular perspective are replaced with a visualization of the non-visible properties from the set of data points.

Abstract: The present disclosure provides an improved approach to implement structure learning of neural networks by exploiting correlations in the data/problem the networks aim to solve. A greedy approach is described that finds bottlenecks of information gain from the bottom convolutional layers all the way to the fully connected layers. Rather than simply making the architecture deeper, additional computation and capacitance is only added where it is required.

Abstract: Methods and systems for videoconferencing include recognition of emotions related to one videoconference participant such as a customer. This ultimately enables another videoconference participant, such as a service provider or supervisor, to handle angry, annoyed, or distressed customers. One example method includes the steps of receiving a video that includes a sequence of images, detecting at least one object of interest (e.g., a face), locating feature reference points of the at least one object of interest, aligning a virtual face mesh to the at least one object of interest based on the feature reference points, finding over the sequence of images at least one deformation of the virtual face mesh that reflect face mimics, determining that the at least one deformation refers to a facial emotion selected from a plurality of reference facial emotions, and generating a communication bearing data associated with the facial emotion.

Abstract: An article recognition device includes an image interface, a distance information interface, a support unit, and a processor. The support unit supports an article at a height equal to or greater than a minimum discrimination distance of a distance sensor from a reference plane of the distance sensor and includes a member of which a distance cannot be measured by the distance sensor. The processor acquires a captured image capturing the article supported by the support unit through the image interface, acquires distance information measured by the distance sensor toward the reference plane through the distance information interface, extracts an area having a height equal to or greater than the minimum discrimination distance from the reference plane as an article area of the article from the captured image based on the distance information, and recognizes the article from an image of the article area.

Abstract: A method of detecting a potato virus in a crop image depicting at least one potato plant includes storing the crop image in a memory, identifying a first region of the crop image depicting potato plant leaves, identifying a plurality of edges within the first region, determining whether an image segment of the crop image within the first region satisfies one or more leaf creasing criteria symptomatic of leaf creasing caused by the virus based on the edges that are located within the image segment, determining whether the image segment satisfies one or more color criteria symptomatic of discoloration caused by the virus, and determining whether the segment displays symptoms of potato virus based on whether the image segment satisfies one or more of the leaf creasing criteria and the color criteria. A system and computer readable medium are also disclosed.

Abstract: Systems and methods for joint reconstruction and segmentation of organs from magnetic resonance imaging (MRI) data are provided. Sparse MRI data is received at a computer system, which jointly processes the MRI data using a plurality of reconstruction and segmentation processes. The MRI data is processed using a joint reconstruction and segmentation process to identify an organ from the MRI data. Additionally, the MRI data is processed using a channel-wise attention network to perform static reconstruction of the organ from the MRI data. Further, the MRI data can is processed using a motion-guided network to perform dynamic reconstruction of the organ from the MRI data. The joint processing allows for rapid static and dynamic reconstruction and segmentation of organs from sparse MRI data, with particular advantage in clinical settings.

Abstract: Provided are a computing apparatus for constructing a mosaic image and an operation method of the same. The computing apparatus includes: a memory storing one or more instructions; and a processor configured to execute the one or more instructions stored in the memory to: segment an input image into a plurality of sub areas to obtain a plurality of sub area images, extract a feature from each of the plurality of sub area images, generate a plurality of source images respectively corresponding to the plurality of sub areas using an image generation neural network, the image generation neural network using, as a condition, the feature extracted from each of the plurality of sub area images, and combine the plurality of source images respectively corresponding to the plurality of sub areas to generate a mosaic image.

Abstract: The present disclosure relates to improvements in systems and methods in acquiring images via imaging devices, where such imaging devices can be configured, in some implementations, with an unmanned aerial vehicle or other vehicle types, as well as being hand-held. Images are acquired from the imaging devices according to capture plans where useful information types about a structure of interest (or objects, items, etc.) can be derived from a structure image acquisition event. Images acquired from capture plans can be evaluated to generate improvements in capture plans for use in subsequent structure image acquisition events. Capture plans provided herein generate accurate information as to all or part of the structure of interest, where accuracy is in relation to the real-life structure incorporated in the acquired images.

Abstract: A computing device accesses a machine learning model trained on training data of first bonding operations (e.g., a ball and/or stitch bond). The first bonding operations comprise operations to bond a first set of multiple wires to a first set of surfaces. The machine learning model is trained by supervised learning. The device receives input data indicating process data generated from measurements of second bonding operations. The second bonding operations comprise operations to bond a second set of multiple wires to a second set of surfaces. The device weights the input data according to the machine learning model. The device generates an anomaly predictor indicating a risk for an anomaly occurrence in the second bonding operations based on weighting the input data according to the machine learning model. The device outputs the anomaly predictor to control the second bonding operations.

Abstract: Implementations generally relate to surgical scene assessment based on computer vision. In some implementations, a method includes receiving a first image frame of a plurality of image frames associated with a surgical scene. The method further includes detecting one or more objects in the first image frame. The method further includes determining one or more positions corresponding to the one or more objects. The method further includes tracking each position of the one or more objects in other image frames of the plurality of image frames.

Abstract: The trajectory tracking method for a mobile electronic device may include tracking a trajectory of the electronic device by using results of pose estimation using odometry and results of pose estimation using visual localization (VL) as camera pose information.

Abstract: An automated kitchen assistant system inspects a food preparation area in the kitchen environment using a novel sensor combination. The combination of sensors includes an Infrared (IR) camera that generates IR image data and at least one secondary sensor that generates secondary image data. The IR image data and secondary image data are processed to obtain combined image data. A trained convolutional neural network is employed to automatically compute an output based on the combined image data. The output includes information about the identity and the location of the food item. The output may further be utilized to command a robotic arm, kitchen worker, or otherwise assist in food preparation. Related methods are also described.

Abstract: An image detection device and an image detection method are provided. The image detection method includes: obtaining an image, where the image includes an object; adjusting a first size of the image to generate an adjusted image; generating a first divided image and a second divided image according to the image; and detecting the object in the image based on a plurality of input images, where the plurality of input images includes the first divided image, the second divided image, and the adjusted image.

Abstract: Systems and methods for video presentation and analytics for live sporting events are disclosed. At least two cameras are used for tracking objects during a live sporting event and generate video feeds to a server processor. The server processor is operable to match the video feeds and create a 3D model of the world based on the video feeds from the at least two cameras. 2D graphics are created from different perspectives based on the 3D model. Statistical data and analytical data related to object movement are produced and displayed on the 2D graphics. The present invention also provides a standard file format for object movement in space over a timeline across multiple sports.

Abstract: Techniques for detecting cycle data can include determining object properties and motion properties in a set of consecutive frames of a sensor stream. The cycle data can be determined from the object properties and motion properties without detecting constituent objects. The object properties and motion properties enable improved detection of cycle data in the presence of different object poses, different positions of the object, partial occlusion of the object, varying illumination, variation in the background, and or the like.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving, by a neural network (NN), a dataset for generating features from the dataset. A first set of features is computed from the dataset using at least a feature layer of the NN. The first set of features i) is characterized by a measure of informativeness; and ii) is computed such that a size of the first set of features is compressible into a second set of features that is smaller in size than the first set of features and that has a same measure of informativeness as the measure of informativeness of the first set of features. The second set of features if generated from the first set of features using a compression method that compresses the first set of features to generate the second set of features.

Abstract: This disclosure provides methods and systems for classifying biological particles, e.g., blood cells, microbes, circulating tumor cells (CTCs). Using impedance flow cytometry, such as multi-frequency impedance cytometry, in conjunction with supervised machine learning, the disclosed methods and systems demonstrated improved accuracy in classifying biological particles.

Abstract: The present disclosure relates to the generation of partial surface models from volumetric datasets for subsequent registration of such partial surface models to surface topology datasets. Specifically, given an object that is imaged using surface topology imaging and another volumetric modality, the volumetric dataset is processed in combination with an approach viewpoint to generate one or more partial surfaces of the object that will be visible to the surface topology imaging system. This procedure can eliminate internal structures from the surfaces generated from volumetric datasets, thus increases the similarity of the dataset between the two different modalities, enabling improved and quicker registration.

Abstract: An imaging device includes one or more processors; and a computer readable medium storing instructions that, when executed by the one or more processors, cause the imaging device to perform functions including: capturing a first image and thereafter a second image; making a determination of whether or not a difference between the first image and the second image is greater than a threshold value; generating a third image by processing the second image using an image processing algorithm that corresponds to the determination; and displaying the third image.