Abstract: A method for manufacturing a LiDAR device is proposed. The method may include providing a LiDAR module including a laser emitting module and a laser detecting module to a target region. The method may also include adjusting, on the basis of first detecting data obtained from the laser detecting module, a relative position of a detecting optic module with respect to the laser detecting module. The method may further include adjusting, on the basis of image data obtained from at least one image sensor, a relative position of an emitting optic module with respect to the laser emitting module.

Abstract: An electronic apparatus includes: an image sensor for acquiring an image including normal pixel values that are sensed during a first exposure time and short exposure pixel values that are sensed during a second exposure time that is shorter than the first exposure time; and a processor configured to acquire flag information that indicates that a selected region, among a plurality of regions of the image, is one of a motion region and a normal region by using an exposure ratio of the first exposure time and the second exposure time and by using a short exposure pixel value and normal pixel values, which are included in the selected region, and configured to output a restoration image including a restoration pixel value that is corrected from the short exposure pixel value that is included in the selected region, based on the flag information.

Abstract: A DNN accelerator system includes a plurality of accelerator nodes each including a plurality of NAND flash memories, a flash memory system (FMS) controller for controlling the plurality of NAND flash memories, and a tensor buffer, and a processor configured to generate an operation sequence of the plurality of accelerator nodes, in which a DNN model is trained in a data parallel manner using the plurality of accelerator nodes.

Abstract: An image processing system includes: an image signal processor including a first neural network, and processing an input image by using the first neural network so as to generate a post-processed image; and a discriminator including a second neural network, and receiving a target image and the post-processed image, and discriminating the target image and the post-processed image into a real image and a fake image by using the second neural network, wherein the second neural network is trained to discriminate the target image as a real image and to discriminate the post-processed image as a fake image, and the first neural network is trained in such a manner that the post-processed image is discriminated as a real image by the second neural network.

Abstract: An electronic apparatus includes: an image sensor for acquiring an image including normal pixel values that are sensed during a first exposure time and short exposure pixel values that are sensed during a second exposure time that is shorter than the first exposure time; and a processor configured to acquire flag information that indicates that a selected region, among a plurality of regions of the image, is one of a motion region and a normal region by using an exposure ratio of the first exposure time and the second exposure time and by using a short exposure pixel value and normal pixel values, which are included in the selected region, and configured to output a restoration image including a restoration pixel value that is corrected from the short exposure pixel value that is included in the selected region, based on the flag information.

Abstract: An image sensing device includes an analysis module suitable for analyzing, based on pixel values of a kernel, an image texture of the kernel including a target pixel group and one or more adjacent pixel groups, a sum module suitable for generating any one of a first target sum value and a second target sum value based on an analysis result of the analysis module, wherein first target sum value is obtained by applying texture characteristics of the kernel in target pixel values of the target pixel group and the second target sum value is obtained without applying the texture characteristics of the kernel in the target pixel values, and a processing module suitable for generating a sum image based on any one of the first and second target sum values.

Abstract: An image sensing device includes an analysis module suitable for analyzing, based on pixel values of a kernel, an image texture of the kernel including a target pixel group and one or more adjacent pixel groups, a sum module suitable for generating any one of a first target sum value and a second target sum value based on an analysis result of the analysis module, wherein first target sum value is obtained by applying texture characteristics of the kernel in target pixel values of the target pixel group and the second target sum value is obtained without applying the texture characteristics of the kernel in the target pixel values, and a processing module suitable for generating a sum image based on any one of the first and second target sum values.

Abstract: An image processing system includes: an image signal processor including a first neural network, and processing an input image by using the first neural network so as to generate a post-processed image; and a discriminator including a second neural network, and receiving a target image and the post-processed image, and discriminating the target image and the post-processed image into a real image and a fake image by using the second neural network, wherein the second neural network is trained to discriminate the target image as a real image and to discriminate the post-processed image as a fake image, and the first neural network is trained in such a manner that the post-processed image is discriminated as a real image by the second neural network.

Abstract: Provided is a history based CU depth determining method. A history based CU depth determining method according to an exemplary embodiment of the present disclosure is a method for determining depths of a plurality of coding units (CU) included in each of a plurality of coding tree units (CTU) which configures a frame of a video, including: dividing a plurality of previous CTUs of a plurality of previous frames which are the same position as a current CTU of a current frame into a plurality of areas to generate depth history information including information of a CU depth for each of the plurality of areas; determining a plurality of depth candidates for a CU depth for each of the plurality of areas of the current CTU, based on the depth history information; and selecting an optimal CU depth among the plurality of depth candidates for each of the plurality of areas of the current CTU, through a rate-distortion cost (RD-cost) calculation.

Abstract: The present invention is to provide a continuous anion polymerization method for preparing transparent high impact vinyl aromatic polymer comprising the steps of: i) charging and polymerizing vinyl aromatic monomers and initiator to obtain vinyl aromatic polymer in the presence of solvent and polar organic compounds in the first reactor (1a); ii) continuously charging and polymerizing vinyl aromatic monomers and conjugated diene monomers to the vinyl aromatic polymer obtained in the first reactor to obtain tapered vinyl aromatic polymer chain in the presence of organic solvent in the second reactor (1b); iii) continuously charging and polymerizing vinyl aromatic monomers and multi-functional compounds to the vinyl aromatic polymer chain obtained in the second reactor to obtain high impact vinyl aromatic polymer in the presence of organic solvent in the third reactor (1c); and iv) terminating the reaction with terminator and removing the remaining hydrocarbon solvent and unreacted monomers to obtain transparen

Abstract: The present invention is to provide a continuous anion polymerization method for preparing transparent high impact vinyl aromatic polymer comprising the steps of: i) charging and polymerizing vinyl aromatic monomers and initiator to obtain vinyl aromatic polymer in the presence of solvent and polar organic compounds in the first reactor (1a); ii) continuously charging and polymerizing vinyl aromatic monomers and conjugated diene monomers to the vinyl aromatic polymer obtained in the first reactor to obtain tapered vinyl aromatic polymer chain in the presence of organic solvent in the second reactor (1b); iii) continuously charging and polymerizing vinyl aromatic monomers and multi-functional compounds to the vinyl aromatic polymer chain obtained in the second reactor to obtain high impact vinyl aromatic polymer in the presence of organic solvent in the third reactor (1c); and iv) terminating the reaction with terminator and removing the remaining hydrocarbon solvent and unreacted monomers to obtain transparen