Abstract: Techniques related to training and implementing convolutional neural networks for object recognition are discussed. Such techniques may include applying, at a first convolutional layer of the convolutional neural network, 3D filters of different spatial sizes to an 3D input image segment to generate multi-scale feature maps such that each feature map has a pathway to fully connected layers of the convolutional neural network, which generate object recognition data corresponding to the 3D input image segment.

Abstract: Techniques are disclosed for text conditioned image searching. A methodology implementing the techniques according to an embodiment includes receiving a source image and a text query defining a target image attribute. The method also includes decomposing the source image into image content and style feature vectors and decomposing the text query into text content and style feature vectors, wherein image style is descriptive of image content and text style is descriptive of text content. The method further includes composing a global content feature vector based on the text content feature vector and the image content feature vector and composing a global style feature vector based on the text style feature vector and the image style feature vector. The method further includes identifying a target image that relates to the global content feature vector and the global style feature vector so that the target image relates to the target image attribute.

Abstract: This patent provides a novel stereoscopic imaging system. In the preferred embodiment, the improved stereoscopic imaging system would be incorporated onto a smart phone, which is called the stereoscopic smart phone (SSP). The SSP would work in conjunction with one or more stereoscopic head display units (SHDUs). Once this system is operational, it will allow significant improvements to obtaining stereoscopic imagery. One novel aspect of this patent comprises wherein the stereoscopic cameras on the SSP can move in position to alter stereo separation distance and change the convergence to optimally image a scene.

Abstract: Methods and systems for determining an orthodontic treatment for a patient are provided. The method comprises: receiving image data associated with a patient's skull; conducting, based on the image data, a cephalometric analysis of the patient's skull; identifying, via the cephalometric analysis, a first pair of reference points; identifying, via the cephalometric analysis, a second pair of reference points; generating based on the first and second pairs of reference points a first reference line and the second reference line, respectively; determining, based on an intersection of the first reference line and the second reference line, a rotation center of a patient's mandible; and based on the rotations center, determining the orthodontic treatment for the patient.

Abstract: The method for property typicality determination can include: determining a property, determining attribute values for the property, determining a reference population for the property, determining reference population attribute values, determining a typicality metric for the property, and optionally determining an influential attribute.

Abstract: Disclosed is a method for processing hyperspectral data for identifying a marine object including configuring a learning dataset for identifying a marine object using hyperspectral data; analyzing the dataset using an analysis algorithm; managing pixels constituting the object using the analysis result; and constructing learning data using pixel data. According to the present invention, according to the accumulation of the hyperspectral data, it is possible to learn and relearn a hyperspectral data-based marine object recognition model that recognizes marine objects shown as noise in machine learning clustering.

Abstract: The present invention relates to a method of detecting an anomaly using a plurality of images. The method comprises determining a subset of images from the plurality of images, comprising one or more first features. Further, one or more regions comprising one or more second features are identified in each image from the subset of images. Finally, detecting the anomaly from the subset of images based on the one or more regions. Therefore, accuracy of detecting anomalies is increased by using the subset of images.

Abstract: A coding and modulation apparatus and method are presented. The apparatus (10) comprises an encoder (11) that encodes input data into cell words, and a modulator (12) that modulates said cell words into constellation values of a non-uniform constellation. The modulator (12) is configured to use, based on the total number M of constellation points of the constellation, the signal-to-noise ratio SNR in dB and the channel characteristics, a non-uniform constellation from a group of constellations comprising one or more of predetermined constellations defined by the constellation position vector u1 . . . v, wherein v=sqrt(M)/2?1.

Abstract: A method for using piecewise forecasts involves obtaining, by a model discovery service, a plurality of models and generating, by a demand prediction service, a plurality of values for a time series variable. The plurality of values corresponding to a plurality of days to be predicted. The method further involves inputting the plurality of values for the time series variable as part of a piecewise forecast to a headcount estimation service and generating, by the headcount estimation service with the piecewise forecast, an estimated headcount from the time series variable.

Abstract: A method for retrieving footprint images is provided, comprising: pre-training models; cleaning footprint data and conducting expansion pre-processing by using the pre-trained models, dividing the footprint data into multiple data sets; adjusting full connection layers and classification layers of the models; training the models again by using the data sets through the parameters of the pre-trained models; saving the models trained twice, removing the classification layer, executing a feature extraction for images in an image library and a retrieval library to form a feature index library; connecting the features extracted by three models to form fused features, establishing a fused feature vector index library; extracting the features of the images in the image library to be retrieved in advance, and establishing a feature vector library; calculating distances in the retrieval library and the image library when a single footprint image is inputted, thereby outputting the image with the highest similarity.

Abstract: In one embodiment, a system receives an image depicting a line of text. The system segments the image into two or more fragment images. For each of the two or more fragment images, the system determines a first hypothesis to segment the fragment image into a first plurality of grapheme images and a first fragmentation confidence score. The system determines a second hypothesis to segment the fragment image into a second plurality of grapheme images and a second fragmentation confidence score. The system determines that the first fragmentation confidence score is greater than the second fragmentation confidence score. The system translates the first plurality of grapheme images defined by the first hypothesis to symbols. The system assembles the symbols of each fragment image to derive the line of text.

Abstract: Various embodiments enable computers to automatically create photographic lineups for police use and, in so doing, eliminate risks associate with subject judgment involved in human selection of fillers for such photographic lineups. Moreover, various improve the reliability of photographic lineups by selecting images of fillers that are similar to, but not too similar to, an image of the suspect.

Abstract: The invention provides artificial intelligence-enabled image recognition methods and systems for continuously training a computer system to accurately identify a surgical item in a tray using at least 100 randomly created 2-dimensional images of a 3-dimensional synthetic item having unique identifiers assigned to the images or item. The invention also provides an artificial intelligence-enabled image recognition method and system for use to determine whether surgical instruments are present or missing on a surgical tray, and, if applicable, identifying those missing. In one aspect, a server receives an image and analyzes the image with a deep convolutional neural network to classify the type of tray and then compares a list of items that should be on the tray to that which the computer recognizes on the tray to generate an output displayed to a user identifying the items present and/or missing.

Abstract: A distributed learning method of a server managing an ID card recognition model includes releasing an ID card recognition model performing at least one convolution operation on an ID card image captured in a user terminal so that the user terminal uses the ID card recognition model, receiving update information of the ID card recognition model generated according to an ID card recognition result of the released ID card recognition model, and verifying the update information received from the user terminal and updating the ID card recognition model using the verified update information.

Abstract: In various embodiments, a prediction application computes a quality score for re-constructed visual content that is derived from visual content. The prediction application generates a frame difference matrix based on two frames included in the re-constructed video content. The prediction application then generates a first entropy matrix based on the frame difference matrix and a first scale. Subsequently, the prediction application computes a first value for a first temporal feature based on the first entropy matrix and a second entropy matrix associated with both the visual content and the first scale. The prediction application computes a quality score for the re-constructed video content based on the first value, a second value for a second temporal feature associated with a second scale, and a machine learning model that is trained using subjective quality scores. The quality score indicates a level of visual quality associated with streamed video content.

Abstract: Disclosed are techniques for determining a distance (or range) between a first wireless entity and a second wireless entity. In an aspect, the first wireless entity transmits a first positioning reference signaling (PRS) signal to the second wireless entity at a first time, where the first PRS signal is received by the second wireless entity at a second time, and receives a second PRS signal from the second wireless entity at a third time, where the second PRS signal is transmitted by the second wireless entity at a fourth time. The first wireless entity enables the distance to be determined by a location computing entity, for example, by a location server, based on the first, second, third, and fourth times. The first wireless entity may be a mobile device or a base station and the second wireless entity may be the other of the mobile device or base station.

Abstract: In various embodiments, a prediction application computes a quality score for re-constructed visual content that is derived from visual content. The prediction application generates a frame difference matrix based on two frames included in the re-constructed video content. The prediction application then generates a first entropy matrix based on the frame difference matrix and a first scale. Subsequently, the prediction application computes a first value for a first temporal feature based on the first entropy matrix and a second entropy matrix associated with both the visual content and the first scale. The prediction application computes a quality score for the re-constructed video content based on the first value, a second value for a second temporal feature associated with a second scale, and a machine learning model that is trained using subjective quality scores. The quality score indicates a level of visual quality associated with streamed video content.

Abstract: Systems, methods, and non-transitory computer-readable media can receive a geometric map and a semantic map associated with a geographic area, the semantic map comprising semantic data associated with vehicle navigation. A first semantic position estimate associated with a first piece of semantic data contained in the semantic map is generated based on semantic data location information associated with the first piece of semantic data. A final position for the first semantic position estimate is received. One or more three-dimensional semantic labels are applied to the geometric map based on the final position of the first semantic position estimate. A warped semantic map is generated. Generating the warped semantic map comprises warping the semantic map based on the one or more three-dimensional semantic labels.

Abstract: A generalizable and interpretable deep learning model for predicting biomarker status and biomarker metrics from histopathology slide images is provided.

Abstract: A method for operating an access node includes generating a configuration message including information specifying a set of reference signals of a first reference signal type and a set of reference signals of a second reference signal type used to identify a new beam, and information specifying random access channel resources allocated for transmitting preamble sequences, wherein each random access channel resource is associated with a reference signal of the first reference signal type, sending, to one or more user equipments (UEs), the configuration message, receiving, from a UE, a preamble sequence on one of the random access channel resources, and determining an identity of the UE in accordance with the preamble sequence and the one of the random access channel resources.