Abstract:
A system and method for decoding quadrature signals includes a quadrature signal generator, a quadrature signal decoder, a key matrix and a driver. The quadrature signal generator generates quadrature signals on rotation. The quadrature signal decoder is configured to convert the quadrature signals into non-overlapping signals. The key matrix is configured to receive the non-overlapping signals. The driver is configured to scan the key matrix to decode the non-overlapping signals to generate an event update corresponding to a direction of rotation of the quadrature signal generator.
Abstract:
A memory bank includes memory cells and an additional cell to determine an operating voltage of the memory bank. The additional cell has an operating margin that is less than a corresponding operating margin of the other memory cells in the memory bank.
Abstract:
A method for decoding a stream encoded using a scalable video coding and including a plurality of layers of frames divided into a plurality of blocks, decodes block-wise in parallel the layers of the stream. A target block in an enhancement layer is decoded as soon as the block data required for its decoding are available from the reference layer.
Abstract:
An on-chip functional debugger includes one or more functional blocks each providing one or more functional outputs. A hierarchical selection tree is formed by one or more selectors having the output of one of the selectors as a final output and individual selector inputs coupled either to a functional output from the functional blocks or to an output of another selector. A selection signal coupled to the select input of each of the selectors to enable a selected one of its output. An output node coupled to the final output. A method of providing on-chip functional debugging is also provided. A desired functional output from one or more available functional outputs is selected and then the selected functional output is coupled to an output node.
Abstract:
The present disclosure relates to a system includes a microcontroller including a neural network, a time-of-flight sensor including a plurality of pixels and configured to perform a capture of a scene comprising a user, the capture comprising, for each pixel, the measurement of a distance from the user and of a signal value. The sensor is further configured to calculate a value of a standard deviation associated with the distance value, and a value of a standard deviation associated with the signal value and a confidence value. The sensor is further configured to provide the values to the neural network. The neural network is configured to generate, based on the values, an estimate of a direction associated with the user. The system further includes a display, the microcontroller is configured to control the display, or another circuit, based on the estimate.
Abstract:
In an embodiment a non-volatile memory device includes a memory plane including at least one memory area including an array of memory cells having two rows and N columns, wherein each memory cell comprises a state transistor having a control gate and a floating gate selectable by a vertical selection transistor buried in a substrate and including a buried selection gate, and wherein each column of memory cells includes a pair of twin memory cells, two selection transistors of the pair of twin memory cells having a common selection gate and a processing device configured to store in the memory area information including a succession of N bits so that, with exception of the last bit of the succession, a current bit of the succession is stored in two memory cells located on the same row and on two adjacent columns and a current bit and the following bit are respectively stored in two twin cells.
Abstract:
According to one aspect, a method for searching, using a computer, for an optimal combination of hyperparameters allows an automatic learning model to be defined. The method includes several hyperparameter combination tests, each hyperparameter combination test including cross-validation, using a validation data set, the cross-validation defining several performance tests, each hyperparameter combination test being stopped if a performance test score is lower than a best score, the cross-validation further including updating the best score when all of the performance scores computed for this cross-validation are higher than the best score, the updated best score then corresponding to the lowest performance score from among the set of performance scores computed for this cross-validation.
Abstract:
A level shifter includes a first input having low voltage switches, and is configured to receive an input voltage and to generate a second voltage which varies between a first low voltage and a first high voltage. The level shifter also includes a first output having medium voltage switches, and is configured to receive the second voltage and to generate a third voltage which varies between the first low voltage and a second high voltage. The level shifter also includes a second input having medium voltage switches, and is configured to receive the third voltage and to generate a fourth voltage which varies between a second low voltage and the second high voltage. The level shifter also includes a second output having medium voltage switches, and is configured to receive the fourth voltage and generate a fifth voltage which varies between the second low voltage and a third high voltage.
Abstract:
According to one aspect, a computer-implemented method compiles a first trained artificial neural network comprising at least one succession of layers including a depthwise convolutional layer, a saturated rectified linear unit layer, and a two-dimensional convolutional layer. The method comprises equalizing between layers, replacing the saturated rectified linear unit layer with an adaptive channel pruning layer to obtain an artificial neural network with a modified topology, tensor quantizing the layers of the artificial neural network with the modified topology, and compiling the quantized artificial neural network with the modified topology.
Abstract:
The present description concerns a method of controlling an analog-to-digital converter, wherein most significant bits are determined by successive approximations implementing a first digital-to-analog converter and a second digital-to-analog converter. Further, least significant bits are determined by a time-to-digital conversion by applying a first ramp to the output of the first converter with a third digital-to-analog converter and by applying a second ramp to the output of the second converter with a fourth digital-to-analog converter. The variation direction of the first and second ramps is determined by the comparison of the outputs of the first and second converters at the end of the successive approximations.