Abstract: A system and a method for recognition of an orchard on a geographic area are provided. The system includes a pre-processing module for deriving a target section of an aerial image containing a parcel of an orchard, an image optimization module for performing customized image processing on the target section of the aerial image, and a recognition module for determining a type and a border of the orchard present on the target section of the aerial image with a deep learning mechanism. Accordingly, farmers and agricultural entities can effectively monitor orchards within different geographic areas so as to yield better fruit production and conduct better fruit quality control and land utilization.

Abstract: The utility usage of a particular individual occupying a residence may give insight into the individual's current cognitive health and/or to enable provision of various services within the facility for the individual, particularly when monitoring patterns in utility usage over time. To enable accurate and non-invasive utility monitoring, a single-point utility sensor may be secured relative to a utility supply line, and generated signals may be utilized to monitor utility usage and to distinguish between utility fixtures. A centralized computing entity may identify frequency characteristics within the generated data, and may automatically generate one or more machine-learning algorithms to distinguish between utility usage events, without requiring substantial user input.

Abstract: An information processing method performed by a computer includes: obtaining a plurality of recognition result candidates in sensing data and a likelihood of each of the plurality of recognition result candidates, the plurality of recognition result candidates and the likelihood being obtained by inputting the sensing data to a model that is trained by machine learning and performs recognition processing; obtaining an indication designating a part to be analyzed in the sensing data; selecting at least one recognition result candidate from the plurality of recognition result candidates, based on (i) a relationship between each of the plurality of recognition result candidates and the part and (ii) the likelihood of each of the plurality of recognition result, candidates; and outputting the at least one recognition result candidate that is selected.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for machine learning for video analysis and feedback. In some implementations, a machine learning model is trained to classify videos into performance level classifications based on characteristics of image data and audio data in the videos. Video data captured by a device of a user following a prompt that the device provides to the user is received. A set of feature values that describe audio and video characteristics of the video data are determined. The set of feature values are provided as input to the trained machine learning model to generate output that classifies the video data with respect to the performance level classifications. A user interface of the device is updated based on the performance level classification for the video data.

Abstract: A hypernetwork training method includes: acquiring a multipath neural subnetwork based on a preconstructed initial hypernetwork; training the multipath neural subnetwork to update a weight parameter of each substructure in the multipath neural subnetwork; synchronizing the weight parameter of each substructure in the multipath neural subnetwork to the preconstructed initial hypernetwork; and determining whether the preconstructed initial hypernetwork converges, and if it is determined that the preconstructed initial hypernetwork does not converge, re-executing the acquiring, the training, the synchronizing, and the determining, to obtain a target hypernetwork.

Abstract: In various examples, live perception from sensors of a vehicle may be leveraged to detect and classify intersection contention areas in an environment of a vehicle in real-time or near real-time. For example, a deep neural network (DNN) may be trained to compute outputs—such as signed distance functions—that may correspond to locations of boundaries delineating intersection contention areas. The signed distance functions may be decoded and/or post-processed to determine instance segmentation masks representing locations and classifications of intersection areas or regions. The locations of the intersections areas or regions may be generated in image-space and converted to world-space coordinates to aid an autonomous or semi-autonomous vehicle in navigating intersections according to rules of the road, traffic priority considerations, and/or the like.

Abstract: One or more images of a sample may be split into image data split according to dyes. The sample has at least two different dyes, in particular fluorescent dyes. The method for splitting the one or more images includes providing the one or more images of the sample and inputting the one or more images into a machine learning system. The method then includes generating the image data split according to dyes from the image or the images, using the machine learning system. The machine learning system removes at least one partial structure of the sample that is present in the image data split according to dyes of more than one dye from the image data of one or more dyes.

Abstract: A computer-implemented method for classifying a reconstruction includes receiving an uncategorized reconstruction and applying a trained classification function configured to classify the uncategorized reconstruction into one of a plurality of categories. The plurality of categories are based on a labeled data-set including a plurality of labeled reconstructions. The trained classification function uses reconstruction-invariant features for classification. The method further includes storing a label indicating a selected one of the plurality of categories for the uncategorized reconstruction.

Abstract: An object recognition apparatus include a storage device and a processor. The storage device stores peripheral information and tolerance information. The tolerance information is information in which the degree of tolerance for the undetected object is represented for each class of the object. The peripheral information is acquired by a sensor device provided in the vehicle. The processor performs object recognition process for recognizing an object around the vehicle. In the object recognition process, the processor identifies the object and its class to be detected based on the peripheral information, and calculates the likelihood that is a parameter representing the probability of detection of the object. Further, the processor calculates a likelihood threshold corresponding to the object based on the tolerance information, and determines whether to output the identification result of the object based on the comparative between the likelihood and the likelihood threshold.

Abstract: Techniques for fabrication of implant material for the reconstruction of fractured eye orbit may include using an image processing system to analyze a set of two-dimensional images representing a three-dimensional scan of a skull of a patient, automatically detect an orbital fracture in the skull based on the set of two-dimensional images, and identify which/both of the two eye orbits containing any orbital fracture. The techniques may further include, for each of the two-dimensional images in which the orbital fracture is detected, determining a region of interest, and extracting the region of interest. The techniques may further include generating a three-dimensional reconstruction model for the fractured eye orbit, and outputting model data for generating an implant mold for the fractured eye orbit.

Abstract: Diagnosis is inferred by using at least one of a plurality of inferencers configured to infer diagnosis from a medical image and by using a medical image as an input into the at least one of the plurality of inferencers, and the inferred diagnosis is represented.

Abstract: A method of processing a neural network, includes generating an integral map for each channel in a first layer of the neural network based on calculating of area sums of pixel values in first output feature maps of channels in the first layer, generating an accumulated integral map by performing an accumulation operation on the integral maps generated for the respective channels, obtaining pre-output feature maps of a second layer, subsequent to the first layer, by performing a convolution operation between input feature maps of the second layer and weight kernels, and removing offsets in the weight kernels to obtain second output feature maps of the second layer by subtracting accumulated values of the accumulated integral map from pixel values of the pre-output feature maps.

Abstract: The present disclosure generally relates to evaluating communication workflows comprised of tasks using machine-learning techniques. More particularly, the present disclosure relates to systems and methods for generating a prediction of a task outcome of a communication workflow, generating a recommendation of one or more tasks to add to a partial communication workflow to complete the communication workflow, and generating a vector representation of a communication workflow.

Abstract: A method of training a neural network (Convolutional Neural Network-CNN) including reduced graphical annotation input is provided. The training method can be used to train a Testing CNN that can be used for determining Hemolysis (H), Icterus (I), and/or Lipemia (L), or Normal (N) of a serum or plasma portion of a test specimen. The training method includes capturing training images of multiple specimen containers including training specimens, generating region proposals of the serum or plasma portions of the training specimens; and selecting the best matches for the location, size and shape of the region proposals for the multiple training specimens. The obtained features (network and weights) from the training CNN can be used in a testing CNN. Quality check modules and testing apparatus adapted to carry out the training method, and characterization methods using abounding box regressor are described, as are other aspects.

Abstract: Aspects of the invention include systems and methods that train a teacher neural network using labeled images to obtain a trained teacher neural network, each pixel of each of the labeled images being assigned a label that indicates one of a set of classifications. A method includes providing a set of unlabeled images to the trained teacher neural network to generate a set of soft-labeled images, each pixel of each of the soft-labeled images being assigned a soft label that indicates one of the set of classifications and an uncertainty value associated with the soft label, and training a student neural network with a subset of the labeled images and the set of soft-labeled images to obtain a trained student neural network. Student-labeled images are obtained from unlabeled images using the trained student neural network.

Abstract: A computing device for handling video content analysis, comprises a preprocessing module, for receiving a first plurality of frames and for determining whether to delete at least one of the first plurality of frames according to an event detection, to generate a second plurality of frames according to the determination for the first plurality of frames; a first deep learning module, for receiving the second plurality of frames and for determining whether to delete at least one of the second plurality of frames according to a plurality of features of the second plurality of frames, to generate a third plurality of frames according to the determination for the second plurality of frames; and a second deep learning module, for receiving the third plurality of frames, to generate a plurality of prediction outputs of the third plurality of frames.

Abstract: An object recognition ingestion system is presented. The object ingestion system captures image data of objects, possibly in an uncontrolled setting. The image data is analyzed to determine if one or more a priori know canonical shape objects match the object represented in the image data. The canonical shape object also includes one or more reference PoVs indicating perspectives from which to analyze objects having the corresponding shape. An object ingestion engine combines the canonical shape object along with the image data to create a model of the object. The engine generates a desirable set of model PoVs from the reference PoVs, and then generates recognition descriptors from each of the model PoVs. The descriptors, image data, model PoVs, or other contextually relevant information are combined into key frame bundles having sufficient information to allow other computing devices to recognize the object at a later time.

Abstract: Methods, systems, and apparatus for operations for processing fingerprint images. An example system includes obtaining an original fingerprint trace image of a fingerprint trace left by a user on a target object; inputting the original fingerprint trace image to a pre-trained fingerprint image processing model that is configured to process the original fingerprint trace image in accordance with pre-trained parameters of the pre-trained fingerprint image processing model; obtaining as output a target fingerprint trace image from the fingerprint image processing model; and using the target fingerprint trace image as a test fingerprint image for performing a test on a fingerprint recognition device, wherein a degree of matching between a fingerprint in the target fingerprint trace image and a real fingerprint that corresponds to a same finger is greater than or equal to a predetermined threshold.

Abstract: A method with image recognition includes: extracting, using a feature extraction layer, feature data from an input image received by an image sensor; and outputting a recognition result of an object appearing in the input image, by applying a fixed mask and a variable mask to the extracted feature data, wherein the variable mask is adjusted in response to the extracted feature data.

Abstract: A display device is provided. The display device includes a display panel including a pixels; a fingerprint recognition sensor including an image sensor disposed under a first surface of the display panel; and a processor to control the display panel and the fingerprint recognition sensor. A portion of display pixels are configured to emit light in a fingerprint recognition mode. The image sensor includes pixels, at least a portion of the pixels are phase detection pixels. The image sensor generates a fingerprint image signal and a fingerprint phase signal based on reflected light received while the portion of display pixels emit light. The main processor is further configured to perform an anti-spoofing operation or a fingerprint authentication operation based on the fingerprint image signal and the fingerprint phase signal.