Abstract: A pooling operation method for a convolutional neural network includes the following steps of: reading multiple new data in at least one current column of a pooling window; performing a first pooling operation with the new data to generate at least a current column pooling result; storing the current column pooling result in a buffer; and performing a second pooling operation with the current column pooling result and at least a preceding column pooling result stored in the buffer to generate a pooling result of the pooling window. The first pooling operation and the second pooling operation are forward max pooling operations.

Abstract: A convolution operation device includes a convolution calculation module, a memory and a buffer device. The convolution calculation module has a plurality of convolution units, and each convolution unit performs a convolution operation according to a filter and a plurality of current data, and leaves a part of the current data after the convolution operation. The buffer device is coupled to the memory and the convolution calculation module for retrieving a plurality of new data from the memory and inputting the new data to each of the convolution units. The new data are not a duplicate of the current data. A convolution operation method is also disclosed.

Abstract: A system and method for computing a sparse neural network having a plurality of output layers, each of which has a neuron value. Processing engines (PEs) each have a local memory for storing neurons for use with different weight values in a following cycle. A multiplexer selects between the input neuron or the output of the memory. Output from the multiplexor is received along with a weight input to a multiplier whose output is directed to an integrator. A decomposition technique performs a network computation through the use of intermediate neurons when the input neuron is larger than the local memory capacity, and provides data reuse by reusing neurons stored in local memory. Neural systems can be implemented using a neural index to address each of multiple PEs and a parallel-serial first-in-first-out (FIFO) to serially store values in main memory.

Abstract: A neural network computing engine having an array of charge-trap-transistor (CTT) elements which are utilized as analog multipliers with all weight values preprogrammed into each CTT element as a CTT threshold voltage, with multiplicator values received from the neural network inference mode. The CTT elements perform computations of a fully connected (FC) neural network with each CTT element representing a neuron. Row resistors for each row of CTT element sum output currents as partial summation results. Counted pulse generators write weight values under control of a pulse generator controller. A sequential analog fabric (SAF) feeds multiple drain voltages in parallel to the CTT array to enable parallel analog computations of neurons. Partial summation results are read by an analog-to-digital converter (ADC).

Abstract: An image depth decoder includes an NIR image buffer, a reference image ring buffer and a pattern matching engine. The NIR image buffer stores an NIR image inputted by a stream. The reference image ring buffer stores a reference image inputted by a stream. The pattern matching engine is coupled to the NIR image buffer and the reference image ring buffer, and performs a depth computation according to the NIR image and the reference image to output at least one depth value.

Abstract: An apparatus for measuring temperature of turbine blades, including: a radiation collection device, a data processing module; a master control unit (MCU); a calibration module; and a motion servo. The radiation collection device includes a scan reflector, a collimator lens, a first dichroic mirror, a first focus lens, a visible and near-infrared (VNIR) detector, a second dichroic mirror, a second focus lens, a short-wave infrared (SWIR) detector, a third focus lens, and a medium-wave infrared (MWIR) detector. The calibration module includes a calibration reflection mirror and a blackbody furnace. The scan reflector, the collimator lens, the first dichroic mirror, the second dichroic mirror, the third focus lens, and the MWIR detector are disposed successively along a first optical axis; the first dichroic mirror, the first focus lens, and the VNIR detector are disposed successively along a second optical axis that is perpendicular to the first optical axis.

Abstract: The present invention discloses a mutant of cyclodextrin glycosyltransferase and belongs to the fields of gene engineering and enzyme engineering. According to the present invention, a mutant having higher disproportionation activity of cyclodextrin glycosyltransferase is obtained by mutating the cyclodextrin glycosyltransferase. The disproportionation activity of enzymes of mutants V6D, S90G, T168A, T171A, T383A, G608A and V6D/S90G/T168A/T171A/T383A/G608A is respectively 1.89 times, 1.21 times, 1.21 times, 1.22 times, 1.32 times, 2.03 times and 3.16 times that of the wild enzyme in shake flask fermentations. The present invention has certain significance for the industrial production of cyclodextrin glycosyltransferase, and improves the application potential of the enzyme in food, medicine and chemical industries.

Abstract: A refrigerator includes a receiving chamber and a drawer movably received in the receiving chamber. The drawer is at a first location when it is closed. The refrigerator further includes a first sensor that is fixed to the drawer or a wall of the receiving chamber. The first sensor is configured to detect when the drawer reaches a second location. The first and second locations are spaced apart by a spacing distance.

Abstract: A refrigerator includes a cabinet, a door, a storage container or movable shelf, and a video recording module. The cabinet includes a storage compartment, a pair of oppositely disposed side walls, and a connecting portion that extends along a lateral direction of the cabinet and that is connected to the pair of side walls, the connecting portion having a front engaging face. The door is configured to cover the storage compartment and is connected to the front engaging face when closed. The storage container or movable shelf is located in the storage compartment and may be at least partially pulled out of the storage compartment. The video recording module or an illumination module is disposed at the connecting portion, faces toward a front of the cabinet, and is inclined downward to record a video of/illuminate the storage container or movable shelf that is at least partially pulled out of the storage compartment.

Abstract: A method of collecting radiation information of a turbine blade, the method including: 1) collecting a radiated light from the surface of the turbine blade, analyzing the radiated light using a spectrometer to calculate compositions and corresponding concentrations of combustion gas; 2) calculating an absorption coefficient of the combustion gas at different concentrations; 3) calculating a total absorption rate of the combustion gas at different radiation wavelengths under different concentrations of component gases; 4) obtaining a relationship between the radiation and a wavelength; 5) finding at least 3 bands with a least gas absorption rate; 6) calculating a distance between a wavelength of a strongest radiation point of the turbine blade and the center wavelength, and selecting three central wavelengths closest to the wavelength with the strongest radiation; and 7) acquiring radiation data of the turbine blade in the windows obtained in 6).

Abstract: An image depth decoder includes an NIR image buffer, a reference image ring buffer and a pattern matching engine. The NIR image buffer stores an NIR image inputted by a stream. The reference image ring buffer stores a reference image inputted by a stream. The pattern matching engine is coupled to the NIR image buffer and the reference image ring buffer, and performs a depth computation according to the NIR image and the reference image to output at least one depth value.

Abstract: A multi-layer artificial neural network having at least one high-speed communication interface and N computational layers is provided. N is an integer larger than 1. The N computational layers are serially connected via the at least one high-speed communication interface. Each of the N computational layers respectively includes a computation circuit and a local memory. The local memory is configured to store input data and learnable parameters for the computation circuit. The computation circuit in the ith computational layer provides its computation results, via the at least one high-speed communication interface, to the local memory in the (i+1)th computational layer as the input data for the computation circuit in the (i+1)th computational layer, wherein i is an integer index ranging from 1 to (N?1).

Abstract: A method for compressing multiple original convolution parameters into a convolution operation chip includes steps of: determining a range of the original convolution parameters; setting an effective bit number for the range; setting a representative value, wherein the representative value is within the range; calculating differential values between the original convolution parameters and the representative value; quantifying the differential values to a minimum effective bit to obtain a plurality of compressed convolution parameters; and transmitting the effective bit number, the representative value and the compressed convolution parameters to the convolution operation chip.

Abstract: A wearable inductive-force sensor, which provides high-sensitivity dynamic measurements of both normal force and shear force, utilizes three spiral planar force sensing coils. These spiral planar coils allow the measurement of shear force in the x and y directions and the measurement of a normal force in the z direction. The force sensor is configured to be mounted in various locations, such as an insole of a shoe, so as to provide real-time force sensing of forces that are applied to a patient's feet as they move. In addition, force-measurement electronics used with the force sensor are configured to use resonance-frequency division signal multiplexing to monitor the response of the force sensing coils, which allows the sensor to have minimal complexity, while still being highly sensitive.

Abstract: A convolution operation device includes a convolution operation module, a memory, a scale control module and a scaling unit. The convolution operation module outputs a plurality of convolution operation results containing fractional parts. The memory is coupled to the convolution operation module for receiving and storing the convolution operation results containing the fractional parts, and outputs a plurality of convolution operation input values containing fractional parts. The scale control module is coupled to the convolution operation module and generates a scaling signal according to a total scale of the convolution operation results containing the fractional parts.

Abstract: A method for compressing multiple original convolution parameters into a convolution operation chip includes steps of: determining a range of the original convolution parameters; setting an effective bit number for the range; setting a representative value, wherein the representative value is within the range; calculating differential values between the original convolution parameters and the representative value; quantifying the differential values to a minimum effective bit to obtain a plurality of compressed convolution parameters; and transmitting the effective bit number, the representative value and the compressed convolution parameters to the convolution operation chip.

Abstract: A low-cost single-layer mobile touch-screen is disclosed, including a highly sensitive capacitive sensing circuit, and a corresponding finger 3D position estimation and gesture recognition algorithms compatible with mobile device platforms and that can be readily programmed into the mobile device's application processor.

Abstract: Described is a system and method for enabling visual search for information. With each selection of an object included in an image, additional images that include visually similar objects are determined and presented to the user.

Abstract: A method of collecting radiation information of a turbine blade, the method including: 1) collecting a radiated light from the surface of the turbine blade, analyzing the radiated light using a spectrometer to calculate compositions and corresponding concentrations of combustion gas; 2) calculating an absorption coefficient of the combustion gas at different concentrations; 3) calculating a total absorption rate of the combustion gas at different radiation wavelengths under different concentrations of component gases; 4) obtaining a relationship between the radiation and a wavelength; 5) finding at least 3 bands with a least gas absorption rate; 6) calculating a distance between a wavelength of a strongest radiation point of the turbine blade and the center wavelength, and selecting three central wavelengths closest to the wavelength with the strongest radiation; and 7) acquiring radiation data of the turbine blade in the windows obtained in 6).

Abstract: A convolution operation method includes the following steps of: performing convolution operations for data inputted in channels, respectively, so as to output a plurality of convolution results; and alternately summing the convolution results of the channels in order so as to output a sum result. A convolution operation device executing the convolution operation method is also disclosed.