Abstract: The present disclosure provides a stitching quality evaluation method and system, and a redundancy reduction method and system for low-altitude UAV remote sensing images, and belongs to the technical field of image processing. The method comprises: firstly acquiring ground images using a UAV under a preset overlap degree to obtain a low-altitude UAV remote sensing image set, then stitching the low-altitude UAV remote sensing image set to obtain a stitched image, and finally performing a quality evaluation on the stitched image using an improved BRISQUE algorithm to obtain an image quality score, which is applicable to quality evaluation of visible images and multispectral images at the same time through the improved BRISQUE algorithm. In addition, the present disclosure further provides an image redundancy reduction method based on the improved BRISQUE algorithm, thereby improving the image stitching efficiency and stitching quality.

Abstract: Hyper-hemispherical images may be combined to generate a rectangular projection of a spherical image having an equatorial stitch line along of a line of lowest distortion in the two images. First and second circular images are received representing respective hyper-hemispherical fields of view. A video processing device may project each circular image to a respective rectangular image by mapping an outer edge of the circular image to a first edge of the rectangular image and mapping a center point of the circular image to a second edge of the first rectangular image. The rectangular images may be stitched together along the edges corresponding to the outer edge of the original circular image.

Abstract: Aspects of the technology described herein relate to techniques for calculating, during imaging, a quality of a sequence of images collected during the imaging. Calculating the quality of the sequence of images may include calculating a probability that a medical professional would use a given image for clinical evaluation and a confidence that an automated analysis segmentation performed on the given image is correct. Techniques described herein also include receiving a trigger to perform an automatic measurement on a sequence of images, calculating a quality of the sequence of images, determining whether the quality of the sequence of images exceeds a threshold quality, and performing the automatic measurement on the sequence of images based on determining that the quality of the sequence of images exceeds the threshold quality.

Abstract: An artificial neural network-based method for selecting a surface type of an object is suitable for selecting a plurality of objects. The artificial neural network-based method for selecting a surface type of an object includes performing surface type identification on a plurality of object images by using a plurality of predictive models to obtain a prediction defect rate of each of the predictive models, wherein the object images correspond to surface types of a part of the objects, and cascading the predictive models according to the respective prediction defect rates of the predictive models into an artificial neural network so as to select the remaining objects.

Abstract: There is provided a method for lossy image or video encoding and transmission, including the steps of receiving an input image at a first computer system, encoding the input image using a first trained neural network to produce a latent representation, performing a quantization process on the latent representation to produce a quantized latent, and transmitting the quantized latent to a second computer system.

Abstract: To reduce the reliance on software for complex computations used in machine sensory perception, a sensory perception accelerator may include a neural network accelerator a linear algebra accelerator. The neural network accelerator may include systolic arrays to perform neural network computation circuits concurrently on image data and audio data. The linear algebra accelerator may include matrix computation circuits operable to perform matrix operations on image data and motion data.

Abstract: A storage system that performs irreversible compression on time-series data using a compressor/decompressor based on machine learning calculates a statistical amount value of each of one or more kinds of statistical amounts based on one or more parameters in relation to original data (time-series data input to a compressor/decompressor) and calculates a statistical amount value of each of the one or more kinds of statistical amounts based on the one or more kinds of parameters in relation to decompressed data (time-series data output from the compressor/decompressor) corresponding to the original data. The machine learning of the compressor/decompressor is performed based on the statistical amount value calculated for each of the one or more kinds of statistical amounts in relation to the original data and the statistical amount value calculated for each of the one or more kinds of statistical amounts in relation to the decompressed data.

Abstract: Various implementations disclosed herein include devices, systems, and methods that obtain a three-dimensional (3D) representation of a physical environment that was generated based on depth data and light intensity image data, generate a 3D bounding box corresponding to an object in the physical environment based on the 3D representation, classify the object based on the 3D bounding box and the 3D semantic data, and display a measurement of the object, where the measurement of the object is determined using one of a plurality of class-specific neural networks selected based on the classifying of the object.

Abstract: A computing system automatically detects, in a sequence of video frames, a video frame region that depicts a scoreboard. The video frames of the sequence depict image elements including (i) scoreboard image elements that are unchanging across the video frames of the sequence and (ii) other image elements that change across the video frames of the sequence. Given this, the computing system (a) receives the sequence, (b) engages in an edge-detection process to detect, in the video frames of the sequence, a set of edges of the depicted image elements, (c) identifies a subset of the detected set of edges based on each edge of the subset being unchanging across the video frames of the sequence, and (d) detects, based on the edges of the identified subset, the video frame region that depicts the scoreboard.

Abstract: The present disclosure is directed towards methods and systems for determining multimodal image edits for a digital image. The systems and methods receive a digital image and analyze the digital image. The systems and methods further generate a feature vector of the digital image, wherein each value of the feature vector represents a respective feature of the digital image. Additionally, based on the feature vector and determined latent variables, the systems and methods generate a plurality of determined image edits for the digital image, which includes determining a plurality of set of potential image attribute values and selecting a plurality of sets of determined image attribute values from the plurality of sets of potential image attribute values wherein each set of determined image attribute values comprises a determined image edit of the plurality of image edits.

Abstract: In various embodiments, a training application generates a perceptual video model. The training application computes a first feature value for a first feature included in a feature vector based on a first color component associated with a first reconstructed training video. The training application also computes a second feature value for a second feature included in the feature vector based on a first brightness component associated with the first reconstructed training video. Subsequently, the training application performs one or more machine learning operations based on the first feature value, the second feature value, and a first subjective quality score for the first reconstructed training video to generate a trained perceptual quality model. The trained perceptual quality model maps a feature value vector for the feature vector to a perceptual quality score.

Abstract: A neural network device includes a decimation unit configured to convert a discrete value of an input signal to a discrete value having a smaller step number than a quantization step number of the input signal on the basis of a predetermined threshold value to generate a decimation signal a modulation unit configured to modulate a discrete value of the decimation signal generated by the decimation unit to generate a modulation signal indicating the discrete value of the decimation signal, and a weighting unit including a neuromorphic element configured to output a weighted signal obtained by weighting the modulation signal through multiplication of the modulation signal generated by the modulation unit by a weight according to a value of a variable characteristic.

Abstract: An image analysis method and a related device are provided. The method includes: obtaining an input matrix of a network layer A, the input matrix of the network layer A obtained based on a target type image; obtaining a target convolution kernel and a target convolution step length corresponding to the network layer A, different network layers corresponding to different convolution step lengths; performing convolution calculation on the input matrix and the target convolution kernel according to the target convolution step length to obtain an output matrix of the network layer A, the output matrix used for representing a plurality of features included in the target type image; determining a target preset operation corresponding to the target type image according to a pre-stored mapping relationship between a type image and a preset operation; and performing the target preset operation according to the plurality of features comprised included in the target type image.

Abstract: One variation of a method for autonomously training an animal includes: loading an autonomous training protocol for the animal onto a training apparatus configured to dispense units of a primary reinforcer responsive to behaviors performed by the animal; during an autonomous training session for the animal, accessing a video feed of a working field near the training apparatus, detecting the animal in the video feed, and executing the first autonomous training protocol; calculating a training score for the autonomous training session based on behaviors performed by the animal during the autonomous training session; selecting a manual training protocol, from a set of manual training protocols, based on the training score; generating a prompt to execute the first manual training protocol with the animal during a first manual training session; and transmitting the prompt to a user associated with the animal.

Abstract: Various disclosed embodiments are directed to inpainting one or more portions of a target image based on merging (or selecting) one or more portions of a warped image with (or from) one or more portions of an inpainting candidate (e.g., via a learning model). This, among other functionality described herein, resolves the inaccuracies of existing image inpainting technologies.

Abstract: Systems and methods for detecting anomaly in video data are provided. The system includes a generator that receives past video frames and extracts spatio-temporal features of the past video frames and generates frames. The generator includes fully convolutional transformer based generative adversarial networks (FCT-GANs). The system includes an image discriminator that discriminates generated frames and real frames. The system also includes a video discriminator that discriminates generated video and real video. The generator trains a fully convolutional transformer network (FCTN) model and determines an anomaly score of at least one test video based on a prediction residual map from the FCTN model.

Abstract: A system of enhancing biometric analysis matching utilizes an image sensor, such as a digital camera, to capture an image of a face of a person. The system may perform image enhancement, such as edge and contrast enhancement, prior to performing face matching. The enhancement may be localized to a given image region based on determined region illumination. The system may perform image processing and analysis comprising face detection, alignment, feature extraction, and recognition. A biometric recognition confidence indicator may be generated using the results of the image enhancement and analysis. At least partly in response to the biometric recognition confidence indicator falling below a threshold enhancing recognition confidence using an image of visual indicia captured using the image sensor.

Abstract: The present disclosure provides a computer-implemented method for translating an image and a computer-implemented method for training an image translation model. In the computer-implemented method for translating an image, an image translation request carrying an original image is obtained. The original image is processed to generate a pre-translated image, a mask image and a deformation parameter. The original image is deformed based on the deformation parameter to obtain a deformed image. The deformed image, the pre-translated image and the mask image are merged to generate a target translated image.

Abstract: Waveforms in power grids typically reveal a certain pattern with specific features and peculiarities driven by the system operating conditions, internal and external uncertainties, etc. This prompts an observation of different types of waveforms at the measurement points (substations). An innovative next-generation smart sensor technology includes a measurement unit embedded with sophisticated analytics for power grid online surveillance and situational awareness. The smart sensor brings additional levels of smartness into the existing phasor measurement units (PMUs) and intelligent electronic devices (IEDs). It unlocks the full potential of advanced signal processing and machine learning for online power grid monitoring in a distributed paradigm.

Abstract: A method, apparatus and a computer program product for automated data slicing based on an Artificial Neural Network (ANN). The method comprising: obtaining an ANN, wherein the ANN is configured to provide a prediction for a data instance, wherein the ANN comprises a set of nodes having interconnections therebetween; determining an attribute vector based on a subset of the nodes of the ANN; determining, based on the attribute vector, a plurality of data slices; obtaining a testing dataset comprising testing data instances; computing, for each data slice, a performance measurement of the ANN over the data slice, wherein said computing is based on an application of the ANN on each testing data instance that is mapped to the data slice; and performing an action based on at least a portion of the performance measurements of the data slices.