Abstract: Aspects of the present disclosure relate to color correction in image processing pipelines. An example method may include receiving first image data corresponding to reference luminance data and reference chrominance data for each of a plurality of pixels, determining that the first image data corresponds to a raw image captured in a dark environment, generating second image data by performing one or more tone mapping operations on the first image data, the second image data corresponding to current luminance data and current chrominance data for each of the plurality of pixels, and generating output image data. For each pixel of the plurality of pixels, the output data may include an output luminance value of a corresponding pixel of the current luminance data, and chrominance values of the corresponding pixel from a selected one of the reference chrominance data and the current chrominance data, the selection based at least in part on the reference chrominance data and the current chrominance data.

Abstract: Disclosed are a bispecific recombinant protein and a use thereof. The bispecific recombinant protein comprises a first functional binding region, a second functional binding region and an Fc region; the first functional binding region comprises a first chain and a second chain, and a C-terminus of the first chain and a C-terminus of the second functional binding region are separately connected to an N-terminus of the Fc region directly or by means of a linker sequence; and a C-terminus of the second chain is connected to an N-terminus of the second functional binding region directly or by means of a linker sequence. Further disclosed is an application of the recombinant protein in terms of anti-tumor, anti-virus, liver disease treatment, and other applications. The recombinant protein may significantly improve the killing efficacy of target cells while significantly reducing toxic side effects caused by binding to non-target cells, and has good safety and good freeze-thaw stability.

Abstract: A wear state online monitoring method and system based on transfer learning and multi-source information fusion are provided. The method includes: separately collecting data of a friction pair to be monitored in a friction and wear testing state and in an on-ship state; performing feature extraction on the data of the friction pair to be monitored; based on the extracted features, performing feature transferring through transfer learning to obtain successfully transferred features; training a random forest model by using the successfully transferred features, and performing wear state monitoring by using the trained random forest model. The application achieves online monitoring of the wear states of the shafting, gearbox, and wear-prone components in marine and aerospace power plants.

Abstract: A GPC3-targeting antibody-interferon ? fusion protein and a use thereof are disclosed in this invention. The fusion protein consists of an A chain and a B chain; wherein the amino acid sequence of the A chain is as shown in SEQ ID NO: 1, and the amino acid sequence of the B chain is as shown in SEQ ID NO: 2. The killing effect on target cells, especially tumor cells, can be significantly improved, and toxic side effects caused by binding to non-target cells are significantly reduced. The fusion protein has better safety, lower production cost, and good freeze-thaw stability.

Abstract: A manufacturing method of a circuit board includes: providing a first double-sided copper laminate including a dielectric layer, a first copper foil layer and a copper plating layer, wherein the dielectric layer, wherein the dielectric layer defines a groove, the copper plating layer includes a first copper plating portion in the groove and a second copper plating portion beside the first copper plating portion. A double-sided circuit substrate including base layer and two first wiring layers is provided, wherein each first wiring layer includes a signal line. Conductive paste blocks are disposed in the base layer and on both sides of the signal line; and a first double-sided copper laminate is stacked on each side of the double-sided circuit substrate, disposing the signal line in the groove. The conductive paste blocks are pressed electrically connect same to the second copper plating portions. The present disclosure further provides a circuit board.

Abstract: In a data processing method performed by a chip comprising a processor and a computing core, the processor receives metadata of first data and metadata of second data. The second data is obtained by performing a first operation on the first data, and memory addresses corresponding to elements at adjacent positions in each row of the second data are discontinuous. The processor compares the metadata of the second data with the metadata of the first data to determine the first operation, and further determines a second operation matching the first operation. The computing core then obtains third data based on the first data and the second operation, wherein memory addresses corresponding to elements at adjacent positions in each row of the third data are continuous.

Abstract: A data processing method and system, and a related device are described. The data processing system includes a processor and a computing core. The method includes the processor receiving metadata of first data and metadata of second data, where the second data is obtained by performing a first operation set on the first data, the first operation set includes at least two first operations, and memory addresses corresponding to elements at adjacent positions in each row of the second data are discontinuous. The processor identifies the first operation set based on the metadata of the second data, and determines a second operation set that matches the first operation set. The computing core obtains third data based on the first data and the second operation set, where memory addresses corresponding to elements at adjacent positions in each row of the third data are continuous.

Abstract: The present invention relates to the field of bio-pharmaceuticals and, in particular, to an anti-Siglec-15 antibody or an antigen binding fragment thereof, a nucleic acid sequence encoding the anti-Siglec-15 antibody of the present invention or an antigen binding fragment thereof, an expression vector containing the nucleic acid sequence, a host cell containing the expression vector, a composition or a bispecific antibody comprising the anti-Siglec-15 antibody of the present invention or an antigen binding fragment thereof, an immune conjugate comprising the anti-Siglec-15 antibody of the present invention or an antigen binding fragment thereof linked to a therapeutic agent, as well as use of the anti-Siglec-15 antibody of the present invention or an antigen binding fragment thereof in the preparation of a medicament for treatment of cancer.

Abstract: In accordance with an embodiment, a model processing method includes obtaining a machine learning model, where the machine learning model includes a flag bit; determining an operator queue of a first operator in the machine learning model, where the first operator is an operator corresponding to the flag bit in the machine learning model; determining a first computational graph of the operator queue; and training the machine learning model based on the first computational graph.

Abstract: For a problem that costs of matching a new data format by a current artificial intelligence system are high, an artificial intelligence system and a corresponding data processing method and apparatus are provided, to perform, based on an artificial intelligence task from a user, format conversion on input data corresponding to the artificial intelligence task, so that a data format used by the artificial intelligence system for processing the artificial intelligence task can be extended with low costs, and the artificial intelligence system can select a proper data format to process the artificial intelligence task.

Abstract: A tensor processing method by a first processor includes obtaining a plurality of first tensors; copying the plurality of first tensors into a second tensor, wherein the plurality of first tensors included in the second tensor occupy consecutive spaces in the second tensor; receiving a first processing instruction from a second processor, wherein the first processing instruction includes a first identifier that indicates the second tensor and a first processing identifier that indicates a first processing operation; and processing the second tensor based on the first processing operation.

Abstract: Aspects of the present disclosure relate to color correction in image processing pipelines. An example method may include receiving first image data corresponding to reference luminance data and reference chrominance data for each of a plurality of pixels, determining that the first image data corresponds to a raw image captured in a dark environment, generating second image data by performing one or more tone mapping operations on the first image data, the second image data corresponding to current luminance data and current chrominance data for each of the plurality of pixels, and generating output image data. For each pixel of the plurality of pixels, the output data may include an output luminance value of a corresponding pixel of the current luminance data, and chrominance values of the corresponding pixel from a selected one of the reference chrominance data and the current chrominance data, the selection based at least in part on the reference chrominance data and the current chrominance data.

Abstract: This application provides a data stream identification method and apparatus and belongs to the field of Internet technologies. The method includes: obtaining packet transmission attribute information of N consecutive packets in a target data stream; generating feature images of the packet transmission attribute information of the N consecutive packets based on the packet transmission attribute information of the N consecutive packets; and inputting the feature images into a pre-trained image classification model, to obtain a target application identifier corresponding to the target data stream. According to this application, accuracy of identifying an application identifier corresponding to a data stream can be improved.

Abstract: An air pressure-machine vision based system for measuring a rheological property of a viscoelastic material includes a machine body, a lifting experiment table system, an air pressure generation control system, an image collection system, and a controlling and information processing system, where the lifting experiment table system, the air pressure generation control system, the image collection system and the controlling and information processing system are mounted on the machine body; the lifting experiment table system includes a lifting table stepping motor, an L-shaped lifting table and a lifting table motor driver, and the lifting table motor driver is connected to the lifting table stepping motor and configured to drive the lifting table stepping motor; and the lifting table stepping motor is connected to the L-shaped lifting table and configured to control lifting of the L-shaped lifting table.

Abstract: An air pressure-machine vision based system for measuring a rheological property of a viscoelastic material includes a machine body, a lifting experiment table system, an air pressure generation control system, an image collection system, and a controlling and information processing system, where the lifting experiment table system, the air pressure generation control system, the image collection system and the controlling and information processing system are mounted on the machine body; the lifting experiment table system includes a lifting table stepping motor, an L-shaped lifting table and a lifting table motor driver, and the lifting table motor driver is connected to the lifting table stepping motor and configured to drive the lifting table stepping motor; and the lifting table stepping motor is connected to the L-shaped lifting table and configured to control lifting of the L-shaped lifting table.

Abstract: A manufacturing method of a circuit board includes: providing a first double-sided copper laminate including a dielectric layer, a first copper foil layer and a copper plating layer wherein the dielectric layer, wherein the dielectric layer defines a groove, the copper plating layer includes a first copper plating portion in the groove and a second copper plating portion beside the first copper plating portion. A double-sided circuit substrate including base layer and two first wiring layers is provided, wherein each first wiring layer includes a signal line. Conductive paste blocks are disposed in the base layer and on both sides of the signal line; and a first double-sided copper laminate is stacked on each side of the double-sided circuit substrate, disposing the signal line in the groove. The conductive paste blocks are pressed electrically connect same to the second copper plating portions. The present disclosure further provides a circuit board.

Abstract: A method for manufacturing an antenna printed circuit board includes the following steps: providing a flexible-rigid printed circuit board comprising a flexible region and a rigid region. The flexible region comprises at least one first connecting pad and at least one second connecting pad arranged opposite sides of the rigid region. At least one antenna module is attached to the first connecting pad through a conductive paste. The antenna module is fixed on the rigid region by a first reflow soldering process. At least one radio frequency integrated circuit structure is attached to the second connecting pad through a conductive paste. The radio frequency integrated circuit is fixed on the rigid region by a second reflow soldering process. A gap between the antenna module and the rigid region and a gap between the radio frequency integrated circuit structure and the rigid region are filled with dielectric materials and cured.