Abstract: Method and apparatus of video encoding video coding for a video encoder or decoder using Neural Network (NN) are disclosed. According to one method, input data or a video bitstream are received for blocks in one or more pictures, which comprise one or more colour components. The residual data, prediction data, reconstructed data, filtered-reconstructed data or a combination thereof is derived for one or more blocks of said one or more pictures. A target signal corresponding to one or more of the about signal types is processed using a NN (Neural Network) and the input of the NN or an output of the NN comprises two or more colour components. According to another method, A target signal corresponding to one or more of the about signal types is processed using a NN and the input of the NN or an output of the NN comprises two or more colour components.

Abstract: A method and apparatus for video coding using a coding mode belonging to a mode group comprising an Intra Block Copy (IBC) mode and an Intra mode are disclosed. According to the present invention, for both IBC and Intra mode, a same default scaling matrix is used to derive the scaling matrix for a current block. In another embodiment, for the current block with block size of M×N or N×M, and M greater than N, a target scaling matrix is derived from an M×M scaling matrix by down-sampling the M×M scaling matrix to an M×N or N×M scaling matrix.

Abstract: A method for performing transform skip mode (TSM) in a video decoder is provided. A video decoder receives data from a bitstream to be decoded as a plurality of video pictures. The video decoder parses the bitstream for a first syntax element in a sequence parameter set (SPS) of a current sequence of video pictures. When the first syntax element indicates that transform skip mode is allowed for the current sequence of video pictures and when transform skip mode is used for a current block in a current picture of the current sequence, the video decoder reconstructs the current block by using quantized residual signals that are not transformed.

Abstract: We describe an antibody or antigen-binding fragment thereof comprising a heavy chain variable region (VH) sequence and a light chain variable region (VL) sequence of a clone selected from: 5C4, 4H1, 1D2, 1C4, 2B5, 1F1, 1H1, 4A5, 1D8, 1D3, 2F1, 3C6, 3D12, 3A9, 1F4, 1H10, 3C12, 4G11, 4A12, 1A12, 3G12, 4F9, 4G4 and 1E1 which is capable of specifically binding to glypican-3 (GPC-3) (GenBank Accession Number: NP_004475.1), or a sequence having at least 75%, 80%, 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity thereto.

Abstract: A power flat no-lead (FN) package is provided. The power FN package includes a die paddle; a die, disposed on the die paddle, operating at a radio frequency; a first lead, disposed by a first side of the die paddle, configured to receive an input signal of the power FN package; and a capacitor, disposed on the first lead; wherein a lead width of the first lead is greater than a half of a first side length of the first side.

Abstract: The invention discloses a pure end-to-end deep reinforcement learning for training car racing game AI bot that uses only the velocity information extracted from screen for both training and testing phases without using any internal state from game environment, such as the car facing angle. The learned AI bot can play better than the average performance of human players. In addition, the reward function is designed to consist only the velocity value, and use Ape-X distributed training framework combined with a variant of Deep Q Network to solve the sparse training signal problem caused by the reward function of an original design. Moreover, limit learner rate method is designed that improves the training efficiency and training performance. The AI bot trained in this way can achieve performance beyond the average human level and reach a level close to professional players.

Abstract: An illumination system and a projection device are provided. The illumination system includes a light source device, a light homogenizing element, and a polarization element. The light source device is configured to provide first, second, third, and first compensation beams. The light homogenizing element is disposed on transmission paths of the first, second, and third beams and the first compensation beam and is configured to convert the first, second, and third beams and the first compensation beam into an illumination beam. The polarization element is disposed on the transmission paths of the first beam and the first compensation beam and located between the light source device and the light homogenizing element. The first beam and the first compensation beam pass through the polarization element to change polarization states of the first beam and the first compensation beam to be the same as polarization states of the second and third beams.

Abstract: The invention provides an ultrasound data processing method for pre-processing signal data in advance of generating ultrasound images. The method seeks to reduce noise through application of coherent persistence to a series of raw ultrasound signal representations representative of the same path or section through a body but at different successive times. A motion compensation procedure including amplitude peak registration and phase alignment is applied to raw echo signal data in advance of application of persistence in order to cohere the signals and thereby limit the introduction of motion induced artifacts.

Abstract: A memory device including a first semiconductor die and a memory cube mounted on and connected with the first semiconductor die is described. The memory cube includes multiple stacked tiers, and each tier of the multiple stacked tiers includes second semiconductor dies laterally wrapped by an encapsulant and a redistribution structure disposed on the second semiconductor dies and the encapsulant. The second semiconductor dies of the multiple stacked tiers are electrically connected with the first semiconductor die through the redistribution structures in the multiple stacked tiers.

Abstract: A system includes an acoustic probe and an acoustic imaging machine. The acoustic probe includes a substrate with first and second principal surfaces, a device insertion port with an opening passing through the substrate from the first principal surface to the second principal surface, and an array of acoustic transducer elements supported by the substrate and disposed around the device insertion port. The acoustic imaging machine may systematically vary the size and/or position of the active acoustic aperture of the probe by providing transmit signals to selected acoustic transducer elements to cause the array to transmit an acoustic probe signal to an area of interest and may record a feedback signal of the transmit signals from an acoustic receiver provided at a distal end of an interventional device passed through the device insertion port into the area of interest to find an active acoustic aperture having optimal acoustic performance.

Abstract: A pushing force-actuated braking device includes an annular housing that houses a brake disc, a braking piston, plural braking elements, and a brake-releasing piston. When only the braking piston is under the action of a fluid, the braking piston applies an axial pushing force to the brake disc such that the brake disc is kept at a braking position jointly by the braking piston and the braking elements. When only the brake-releasing piston is under the action of a fluid, the brake-releasing piston applies an opposite pushing force to the brake disc to keep it at a brake-releasing position. Should the fluid acting on the braking piston fail, the force of the braking elements still enables the brake disc to produce a braking effect. The pushing force-actuated braking device has a modular design to facilitate assembly and disassembly. A rotary table using the braking device is also provided.

Abstract: A work function metal gate device includes a gate, a drift region, a source, a drain and a first isolation structure. The gate includes a convex stair-shaped work function metal stack or a concave stair-shaped work function metal stack disposed on a substrate. The drift region is disposed in the substrate below a part of the gate. The source is located in the substrate and the drain is located in the drift region beside the gate. The first isolation structure is disposed in the drift region between the gate and the drain.

Abstract: Exemplary video processing methods and apparatuses for encoding or decoding a current block by inter prediction are disclosed. Input data of a current block is received and partitioned into sub-partitions and motion refinement is independently performed on each sub-partition. A reference block for each sub-partition is obtained from one or more reference pictures according to an initial motion vector (MV). A refined MV for each sub-partition is derived by searching around the initial MV with N-pixel refinement. One or more boundary pixels of the reference block for a sub-partition is padded for motion compensation of the sub-partition. A final predictor for the current block is generated by performing motion compensation for each sub-partition according to its refined MV. The current block is then encoded or decoded according to the final predictor.

Abstract: A temperature calibration method of an infrared thermal image camera and a calibration method of a temperature sensing system of an infrared thermal image camera are provided. The temperature calibration method of the infrared thermal image camera includes: providing a temperature sensing system to measure an object under test, so as to obtain an uncalibrated temperature; and calibrating the uncalibrated temperature to obtain a calibrated temperature, according to the uncalibrated temperature and a temperature calibration function. The temperature calibration function is obtained through a regression analysis.

Abstract: An augmented reality screen system includes an augmented reality device and a host. The augmented reality device is configured to take a physical mark through a camera. The host is configured to receive the physical mark, determine position information and rotation information of the physical mark, and fetch a virtual image from a storage device through a processor of the host. The processor transmits an adjusted virtual image to the augmented reality device according to the position information and the rotation information, and the augmentation device projects the adjusted virtual image to a display of the augmented reality device. The adjusted virtual image becomes a virtual extended screen, and the virtual extended screen and the physical mark are simultaneously displayed on the display of the augmented reality device.

Abstract: Method and apparatus for constrained de-blocking filter are disclosed. According to one method, if a CU is partitioned into multiple sub-CUs, the de-blocking process is also applied to the sub-block boundaries inside the current filtered-reconstructed block. According to another method, if first reference samples used for the de-blocking process of a first boundary are to be modified by the de-blocking process of a second boundary, the first reference samples are replaced by padding samples that are not to be modified by the de-blocking process of the second boundary. According to yet another method, the de-blocking process is applied to a reconstructed block corresponding to a current block to result in a current filtered-reconstructed block regardless whether a boundary of the current block corresponds to an 8×8 sample grid boundaries.

Abstract: A light source module and a projection apparatus comprising the same are provided. The light source module comprises a light source unit, a light splitting element, and a condenser lens. The light source unit is configured to provide first color light beams including a first sub-light beam and a third sub-light beam, and second color light beams including a second sub-light beam and a fourth sub-light beam. The light source unit comprises a first light source unit to provide the first and second sub-light beams, and a second light source unit to provide the third and fourth sub-light beams. One of the first sub-light beam and the third sub-light beam is transmitted to the condenser lens after being reflected by the light splitting element, and the other of the first sub-light beam and the third sub-light beam is transmitted to the condenser lens after passing through the light splitting element.

Abstract: A log processing device and a log processing method thereof are provided. The log processing device divides the original log data into a plurality of block data, transforms a numeric variable of each of the block data into a representative code, and determines whether to perform a combination process for continuous block data to generate a plurality of combinational block data according to a data integrity of each of the block data. The log processing data takes the combinational block data as a log template, and each of the combinational block data corresponds to an event.