Abstract: Disclosed is a method for manufacturing a laminated piezoelectric composite. The method includes wet-mixing ceramic powder, a polymer binder, a plasticizer and a solvent for 4 to 72 hours so as to generate a mixed slurry, introducing the mixed slurry into a tape casting process so as to prepare a plurality of piezoelectric composite sheets, drying and forming the plurality of piezoelectric composite sheets using a roll-to-roll process so as to prepare the plurality of formed piezoelectric composite sheets, forming internal electrodes on the plurality of piezoelectric composite sheets so as to prepare the plurality of piezoelectric composite sheets having the internal electrodes, laminating and pressing the plurality of piezoelectric composite sheets having the internal electrodes so as to generate a piezoelectric composite sheet laminate having the internal electrodes, and cutting the piezoelectric composite sheet laminate having the internal electrodes into a desired shape and size.

Abstract: An image region segmentation method and system suing self-spatial adaptive normalization is provided. The image region segmentation system includes: an encoder configured to encode an image for segmenting a region by using a plurality of encoding blocks; and a decoder configured to decode the image encoded by the encoder and to generate a region-segmented image by using a plurality of decoding blocks, wherein each of the encoding blocks processes an inputted image into a convolution layer, performs spatial adaptive normalization, and then reduces the image and delivers the image to the next encoding block. Accordingly, spatial characteristics of the image are considered in an encoding process and a decoding process, so that region segmentation can be exactly performed with respect to various images.

Abstract: There are provided a method and a system for automatically generating a back-channel in an interactive agent system. According to an embodiment of the disclosure, an automatic back-channel generation method includes: predicting a back-channel by analyzing an utterance of a user inputted in a back-channel prediction model; and generating the predicted back-channel, and the back-channel prediction model is an AI model that is trained to predict a back-channel to express from the utterance of the user. Accordingly, a back-channel is automatically generated by utilizing a back-channel prediction module which is based on a language model, so that a natural dialogue interaction with a user may be implemented in an interactive agent system, and quality of a dialogue service provided to a user may be enhanced.

Abstract: In one embodiment, an integrated light transmission/reception optical system module includes a light receiving lens, a light source, and a light transmitting mirror. The light receiving lens receives light, concentrates the received light on a light detector disposed at a rear position, and has an optical path groove formed to be directed from a circumference to a central portion and formed to expose a front side. The light source outputs a pulse laser along the optical path groove from the circumference of the light receiving lens toward the central portion of the light receiving lens. The light transmitting mirror is disposed within the optical path groove, is located on a path of the pulse laser, and reflects in a front direction the pulse laser outputted from the light source. Other embodiments are also possible.

Abstract: Provided are a method and an apparatus for printing a hologram by using a mask. The hologram printing method according to an embodiment generates a hologram fringe pattern, splits the hologram fringe pattern on a hogel basis, generates the split hogels, masks a part of the generated hogel, and prints the masked hogel on a hologram medium. Accordingly, an empty space which occurs between hogels when a hologram is printed is prevented from being generated by using a mask, so that a fill factor can be effectively enhanced, and eventually, image quality of a hologram can be enhanced.

Abstract: This application relates to a computing system and method for an automated program error repair. In one aspect, the computing system includes a storage, a preprocessing processor, and an automated error repair processor. The storage stores a program code. The preprocessing processor acquires the program code from the storage and preprocesses the program code. Preprocessing includes tokenizing the program code with tokens, converting the tokens into vectors, and adding location information for the tokens. The automated error repair processor receives the preprocessed program code as an input from the preprocessing processor, detects an error in the preprocessed program code, corrects the detected error, and outputs the error-corrected program code. Detecting and correcting the error are performed based on a deep learning result and the location information for the tokens.

Abstract: Provided is a method in which, when configuring an HUD that produces a holographic image at a distance using a holographic optical element (HOE), an HOE capable of correcting aberrations generated by a projection optical system is manufactured and used to improve the quality of an HUD image. A method for manufacturing an HOE for HUD according to an embodiment of the present invention comprises the steps of: measuring aberrations generated by an optical system that projects an image of a display device; recording the measured aberrations in a master HOE; reproducing an aberrated wavefront of the optical system by playing the master HOE on a display plane on which the image of the display device is expressed; and causing an interference of the reproduced aberrated wavefront and a spherical wavefront irradiated from the HUD image plane on which the image of the display device is created, and recording the interference in the HOE.

Abstract: There is provided a cloud management method and apparatus for performing load balancing so as to make a service in a cluster that is geographically close in an associative container environment and has a good resource current status. The cloud management method according to an embodiment includes: monitoring, by a cloud management apparatus, available resource current statuses of a plurality of clusters, and selecting a cluster that owns a first service supported by a first cluster an available resource rate of which is less than a threshold value; calculating, by the cloud management apparatus, scores regarding an available resource current status and geographical proximity of each cluster; and performing, by the cloud management apparatus, load balancing of the first service, based on a result of calculating the scores.

Abstract: Provided is a 5G dual port beamforming antenna. The beamforming antenna according to an embodiment of the present disclosure includes a plurality of patch antennas, and the patch antenna includes: a first patch positioned on an upper portion; a second patch positioned on a left side under the first patch and having a plurality of feeding ports formed thereon; a third patch positioned on an upper portion of a right side of the second patch; and a fourth patch positioned on a lower portion of the right side of the second patch. Accordingly, 3D wide angle beamforming is possible in an antenna to be used in 5G mobile communication systems, military radar systems, etc.

Abstract: This application relates to an aluminum-based amorphous metal particles, a conductive Ink and OLED cathode including the aluminum-based amorphous metal particles, and a method of manufacturing the aluminum-based amorphous metal particles. In one aspect, the amorphous metal particles are represented by a formula AlxLiyNizYwCov. Here, x, y, z, w, and v denote an atomic ratio, and satisfy the following relationships: 75.0?x?90.0, 3.0<y?7.0, 1.0?z?7.0, 2.0?w?10.0, 0.0?v?5.5, and x+y+z+w+v=100.

Abstract: Disclosed herein are a method and apparatus for encoding/decoding a deep learning network. According to an embodiment, the method for decoding a deep learning network may include decoding network header information regarding the deep learning network; decoding layer header information regarding a plurality of layers in the deep learning network; decoding layer data information regarding specific information of the plurality of layers; and obtaining the deep learning network and a plurality of layers in the deep learning network, and the layer header information includes layer distinction information associated with distinguishing the plurality of layers.

Abstract: Disclosed herein are a method and apparatus for encoding/decoding a deep neural network. According to the present disclosure, the method for decoding a deep neural network may include: in a plurality of layers of the deep neural network, entropy decoding quantization information for a current layer; performing dequantization on the current layer; and obtaining a plurality of layers of the deep neural network. At least one of global quantization and local quantization is performed on the current layer.

Abstract: Provided is a method for monitoring 3D printing equipped with a 3D printing slicer and a recursive loop structure. A 3D printing method according to an embodiment of the present invention sets up a slicing environment for 3D printing of a 3D model, generates a mechanical code by performing slicing according to the setup environment, monitors the status of the 3D printing according to the generated mechanical code, and, depending on the monitoring result, determines whether or not to re-perform the setup and subsequent steps. Accordingly, by semi- or fully automating the 3D printing engineering process, the time and effort for engineering performance involving human participation are reduced, and the human resource is concentrated on a more important area, such that the effects of enhancing the 3D printing output quality and assuring the quality can be expected.

Abstract: Provided is a technology for implementing an AR optical waveguide display capable of showing a hologram image by means of a small and simple system configuration by using an HOE. A holographic display according to an embodiment of the present invention comprises: a light source module for emitting a beam; an optical waveguide through which the emitted beam is incident and propagated; a plurality of holographic optical elements (HOES) for propagating the beam incident to the optical waveguide inside the optical waveguide while totally reflecting the beam; and a modulator for reproducing a holographic image through the progressing beam and propagating the beam to the inside of the optical waveguide while totally reflecting the beam. Accordingly, it is possible to implement, as a small and simple system, an optical waveguide display showing an AR hologram by using an optical waveguide and an HOE.

Abstract: There are provided AI model learning method and system based on self-learning for focusing on specific areas. According to an embodiment, a network learning system includes: a detection module configured to detect a specific area from unlabeled images, and to generate unlabeled area images; a configuration module configured to configure self-learning data by using the generated area images; and a learning module to cause a backbone network to perform self-learning by using the configured self-learning data. Accordingly, an AI model may be trained based on self-learning for focusing on a desired specific area according to a desired purpose, and high-performance analysis specified for various purposes and characteristics of various types of specific areas is possible.

Abstract: This application relates to a mechanical button apparatus that includes a body where an upper portion thereof is closed and a lower portion thereof is opened and a bottom plate attached on a display screen of a display apparatus to cover the opened lower portion. The body includes an aperture member including a wing set expanded or contracted in a rotation direction of the body. The body also includes a photovoltaic (PV) cell array substrate including a PV cell configured to generate power for charging a battery and an optical sensor configured to convert light signals, reflected by the expanded or contracted wing set, into electrical signals. The body further includes a circuit substrate configured to operate with the power charged into the battery, calculate number of rotation manipulations of the body by using the electrical signals, and transmit the calculated number of rotation manipulations to the display apparatus.

Abstract: Provided is a slicing 2D data-based pattern application method for generating an output code in which the inside of a model is filled with a pattern, so as to reduce a binder usage amount and maintain a strength and a shape in sand binder jetting additive manufacturing. A slicing 2D data-based pattern application method according to an embodiment of the present invention comprises the steps of: generating 2D data by slicing an output model; generating an inner pattern in at least one of layers forming an output model in consideration of the set thickness of the layers and the outer thickness thereof and generating an output code by applying the generated inner pattern. Therefore, the cost of producing an additive manufacturing output can be reduced by reducing the binder usage amount through the application of the inner pattern.

Abstract: There are provided a method for segmenting a sign language video by gloss to recognize a sign language sentence, and a method for training. According to an embodiment, a sign language video segmentation method receives an input of a sign language sentence video, and segments the inputted sign language sentence video by gloss. Accordingly, there is suggested a method for segmenting a sign language sentence video by gloss, analyzing various gloss sequences from the linguistic perspective, understanding meanings robustly in spite of various changes in sentences, and translating sign language into appropriate Korean sentences.

Abstract: Provided is a nozzle clogging defect compensating method for compensating for a nozzle clogging defect appearing in a binder jetting stack manufacturing means. The nozzle clogging defect compensating method according to an embodiment of the present invention comprises the steps of: determining a defect occurrence region when a clogging defect of a nozzle used in the binder jetting stack manufacturing means occurs; determining whether compensation for the detect occurrence region is possible; when the compensation is possible, generating defect information and reflecting the defect information in an output code; setting a defect compensation region on the basis of the defect information; determining a defect compensation type of the defect compensation region; and reflecting a result of the determining in the output code.

Abstract: There are provided a method and a system for recognizing a finger language video in units of syllables based on AI. The finger language video recognition system includes: an extraction unit configured to extract posture information of a speaker from a finger language video; and a recognition unit configured to recognize a finger language of the speaker from the extracted posture information of the speaker in units of syllables, and to output a text. Accordingly, a language text in units of syllables may be generated from a finger language video, by using an AI-based syllable unit finger language recognition model.