Abstract: A classification device receives sensor data from a set of sensors and generates, using a context classifier having a set of classifier model parameters, a set of raw predictions based on the received sensor data. Temporal filtering and heuristic filtering are applied to the raw predictions, producing filtered predictions. A prediction error is generated from the filtered predictions, and model parameters of the set of classifier model parameters are updated based on said prediction error. The classification device may be a wearable device.

Abstract: A device includes a memory and processing circuitry coupled to the memory. The processing circuitry, in operation, generates an indication of a predicted difference in a direction of arrival (DoA) of a signal using a trained autoregressive model. A predicted indication of a DoA of the signal is generated based on a previous indication of the DoA of the signal and the indication of the predicted difference in the DoA of the signal. The processing circuitry actuates or controls an antenna array based on predicted indications of the DoA of the signal.

Abstract: A neural network classifies an input signal. For example, an accelerometer signal may be classified to detect human activity. In a first convolutional layer, two-valued weights are applied to the input signal. In a first two-valued function layer coupled at input to an output of the first convolutional layer, a two-valued function is applied. In a second convolutional layer coupled at input to an output of the first two-valued functional layer, weights of the second convolutional layer are applied. In a fully-connected layer coupled at input to an output of the second convolutional layer, two-valued weights of the fully connected layer are applied. In a second two-valued function layer coupled at input to an output of the fully connected layer, a two-valued function of the second two-valued function layer is applied. A classifier classifies the input signal based on an output signal of second two-valued function layer.

Abstract: An image processing system has one or more memories and image processing circuitry coupled to the one or more memories. The image processing circuitry, in operation, compares a first image to feature data in a comparison image space using a matching model. The comparing includes: unwarping keypoints in keypoint data of the first image; and comparing the unwarped keypoints and descriptor data associated with the first image to the feature data of the comparison image. The image processing circuitry determines whether the first image matches the comparison image based on the comparing.

Abstract: Methods, microprocessors, and systems are provided for implementing an artificial neural network. Data buffers in virtual memory are coupled to respective processing layers in the artificial neural network. An ordered visiting sequence of layers of the artificial neural network is obtained. A virtual memory allocation schedule is produced as a function of the ordered visiting sequence of layers of the artificial neural network, the schedule including a set of instructions for memory allocation and deallocation operations applicable to the data buffers. A physical memory configuration dataset is computed as a function of the virtual memory allocation schedule for the artificial neural network, the dataset including sizes and addresses of physical memory locations for the artificial neural network.

Abstract: Methods, microprocessors, and systems are provided for implementing an artificial neural network. Data buffers in virtual memory are coupled to respective processing layers in the artificial neural network. An ordered visiting sequence of layers of the artificial neural network is obtained. A virtual memory allocation schedule is produced as a function of the ordered visiting sequence of layers of the artificial neural network, the schedule including a set of instructions for memory allocation and deallocation operations applicable to the data buffers. A physical memory configuration dataset is computed as a function of the virtual memory allocation schedule for the artificial neural network, the dataset including sizes and addresses of physical memory locations for the artificial neural network.

Abstract: An image processing system has one or more memories and image processing circuitry coupled to the one or more memories. The image processing circuitry, in operation, compares a first image to feature data in a comparison image space using a matching model. The comparing includes: unwarping keypoints in keypoint data of the first image; and comparing the unwarped keypoints and descriptor data associated with the first image to the feature data of the comparison image. The image processing circuitry determines whether the first image matches the comparison image based on the comparing.

Abstract: Local descriptors are extracted from digital image information and digital depth information related to digital images. The local descriptors convey appearance description information and shape description information related to the digital images. Global representations of the one or more digital images are generated based on the extracted local descriptors, and are hashed. Visual search queries are generated based on the hashed global representations. The visual search queries include fused appearance description information and shape description information conveyed in the local descriptors. The fusing may occur before the global representations are generated, before the hashing or after the hashing.

Abstract: Apparatus and methods to unwarp at least portions of distorted, electronically-captured images are described. Keypoints, instead of an entire image, may be unwarped and used in various machine-vision algorithms, such as object recognition, image matching, and 3D reconstruction algorithms. When using unwarped keypoints, the machine-vision algorithms may perform reliably irrespective of distortions that may be introduced by one or more image capture systems.

Abstract: Digital image processing circuitry clusters a set of images into a set of first clusters of images and a set of unclustered images. The set of first clusters are merged, generating a set of second clusters of images. Images in the set of unclustered images are assigned to one of a cluster of the set of second clusters of images and an outlier image cluster. The clustered images may be partitioned into subclusters based on detection of objects in the images.

Abstract: First and second video frames in a flow of digital video frames are encoded by extracting respective sets of keypoints and descriptors, each descriptor including a plurality of orientation histograms regarding a patch of pixels centered on the respective keypoint. Once a pair of linked descriptors has been identified, one for each frame, which have a minimum distance from among the distances between any one of the descriptors of the first frame and any one of the descriptors of the second frame, the differences of the histograms of the descriptors linked in the pair are calculated, and the descriptors linked in the pair are encoded as the set including one of the linked descriptors and the histogram differences by subjecting the histogram differences to a thesholding setting at zero all the differences below a certain threshold, to quantization, and to run-length encoding.

Abstract: A laserbeam light source is controlled to avoid light sensitive regions around the laserbeam light source. One or more laserlight-sensitive regions are identified based on images of an area around the laserbeam light source, and indications of positions corresponding to the laserlight-sensitive regions are generated. The laserbeam light source is controlled based on the indications of the positions. The laserbeam light source may be controlled to deflect a laserlight beam away from laserlight-sensitive regions, to reduce an intensity of a laserlight beam directed towards a laserlight-sensitive region, etc. Motion estimation may be used to generate the indications of positions corresponding to the laserlight-sensitive regions.

Abstract: A neural network classifies an input signal. For example, an accelerometer signal may be classified to detect human activity. In a first convolutional layer, two-valued weights are applied to the input signal. In a first two-valued function layer coupled at input to an output of the first convolutional layer, a two-valued function is applied. In a second convolutional layer coupled at input to an output of the first two-valued functional layer, weights of the second convolutional layer are applied. In a fully-connected layer coupled at input to an output of the second convolutional layer, two-valued weights of the fully connected layer are applied. In a second two-valued function layer coupled at input to an output of the fully connected layer, a two-valued function of the second two-valued function layer is applied. A classifier classifies the input signal based on an output signal of second two-valued function layer.

Abstract: A neural network includes one layer of neurons including neurons having neuron connections to neurons in the layer and input connections to a network input. The neuron connections and the input connections have respective neuron connection weights and input connection weights. The neurons have neuron responses set by an activation function with activation values and include activation function computing circuits configured for computing current activation values of the activation function as a function of previous activation values of the activation function and current network input values.

Abstract: A laserbeam light source is controlled to avoid light sensitive regions around the laserbeam light source. One or more laserlight-sensitive regions are identified based on images of an area around the laserbeam light source, and indications of positions corresponding to the laserlight-sensitive regions are generated. The laserbeam light source is controlled based on the indications of the positions. The laserbeam light source may be controlled to deflect a laserlight beam away from laserlight-sensitive regions, to reduce an intensity of a laserlight beam directed towards a laserlight-sensitive region, etc. Motion estimation may be used to generate the indications of positions corresponding to the laserlight-sensitive regions.

Abstract: A classification device receives sensor data from a set of sensors and generates, using a context classifier having a set of classifier model parameters, a set of raw predictions based on the received sensor data. Temporal filtering and heuristic filtering are applied to the raw predictions, producing filtered predictions. A prediction error is generated from the filtered predictions, and model parameters of the set of classifier model parameters are updated based on said prediction error. The classification device may be a wearable device.

Abstract: In an embodiment, digital video frames in a flow are subjected to a method of extraction of features including the operations of: extracting from the video frames respective sequences of pairs of keypoints/descriptors limiting to a threshold value the number of pairs extracted for each frame; sending the sequences extracted from an extractor module to a server for processing with a bitrate value variable in time; receiving the aforesaid bitrate value variable in time at the extractor as target bitrate for extraction; and limiting the number of pairs extracted by the extractor to a threshold value variable in time as a function of the target bitrate.

Abstract: Human activities are classified based on activity-related data and an activity-classification model trained using a classification-equalized training data set. A classification signal is generated based on the classifications. The classification-equalized training data set, may, for example, includes a first class having a first sequence length and a number of samples N, and one or more additional classes each having a respective sequence length tj and a respective number of samples Nj determined based on the number of samples N of the first class. For example, a respective sequence length tj and a respective number of samples Nj which satisfy: (i) Nj>N, for sequence length tj; and (ii) Nj<N, for tj?1. The activity-related data may include one or more of acceleration data, orientation data, position data, and physiological data.

Abstract: Searches performed in a data base using image descriptors of query images are managed via a mobile communication device, such as a smartphone, a tablet, etc., by: extracting at the mobile device grayscale and color descriptors of query images, sending the grayscale descriptors as compressed grayscale descriptors of query images from the mobile device to a server for searching in the data base, receiving at the mobile device results of the search, using color descriptors of query images in disambiguating the results by: i) sending the color descriptors as compressed color descriptors of query images from the mobile device to the server and receiving at the mobile device disambiguated search results from the server, or ii) receiving at the mobile device non-disambiguated search results from the server and disambiguating the search results by means of the color descriptors extracted at the mobile device to produce disambiguated search results.

Abstract: An image processing system includes a first processor that acquires frames of image data. For each frame of data, the first processor generates a Gaussian pyramid for the frame of data, extract histogram of oriented gradient (HOG) descriptors for each level of the Gaussian pyramid, compresses the HOG descriptors, and sends the compressed HOG descriptors. A second processor is coupled to the first processor and is configured to receive the compressed HOG descriptors, aggregate the compressed HOG descriptors into windows, compare data of each window to at least one stored model, and generate output based upon the comparison.