Abstract: The present invention relates to a three-dimensional reconstructing method of a 2D medical image. A three-dimensional reconstructing device includes: a communicator for receiving sequential 2D images with an arbitrary slice gap; a sliced image generator for generating at least one sliced image positioned between the 2D images based on a feature point of the adjacent 2D images; and a controller for reconstructing the 2D image into a 3D image by use of the generated sliced image and providing the 3D image.

Abstract: Provided is a mandibular advancement system including: an upper teeth seating part on which upper teeth of a user is placed; a lower teeth seating part on which lower teeth of the user is placed; a driver connected to the upper teeth seating part and the lower teeth seating part and configured to change a relative location of the lower teeth seating part with respect to the upper teeth seating part; a detector configured to detect biometric information of the user; and a controller configured to control driving of the driver based on the biometric information provided by the detector, wherein the driver is disposed outside an oral cavity of the user.

Abstract: Provided are a method for generating a summary and a system therefor. The method according to some embodiments may include calculating a likelihood loss for a summary model using a first text sample and a first summary sentence corresponding to the first text sample, calculating an unlikelihood loss for the summary model using a second text sample and the first summary sentence, the second text sample being a negative sample generated from the first text sample, and updating the summary model based on the likelihood loss and the unlikelihood loss.

Abstract: A frequency multiplier includes a first digitally controlled delay line (DCDL) configured to receive a first clock signal and generate a second clock signal by changing a phase of the first clock signal, a multiplying delay-locked loop (MDLL) configured to generate a third clock signal by multiplying a frequency of the second clock signal, and a DCDL calibration circuit configured to receive the second clock signal and generate a gain signal for adjusting a gain of the first DCDL where a difference between a maximum delay and a minimum delay of the first DCDL is substantially equal to a period of the third clock signal.

Abstract: A binder solution for an all-solid-state battery, an electrode slurry for an all-solid-state battery including the same and a method of manufacturing an all-solid-state battery using the same, and more particularly to a binder solution for an all-solid-state battery, in which a polymer binder configured such that a non-polar functional group is bonded to the end of a polar functional group is used, whereby the polar functional group is provided by a deprotection mechanism of the polymer binder through a thermal treatment, thus increasing adhesion between electrode materials to thereby improve battery capacity and enabling a wet process to thereby reduce manufacturing costs, an electrode slurry for an all-solid-state battery including the same and a method of manufacturing an all-solid-state battery using the same.

Abstract: A method of controlling injection of a drug includes receiving a captured medical image of an object, identifying an object of interest from the medical image, based on a position and a shape of the object of interest, determining a drug injection position, based on the position and the shape of the object of interest, determining a plurality of electrode insertion positions, the plurality of electrodes being configured to supply a current to move the drug, defining a drug movement path based on a shape of the identified object of interest, generating current sequence information to move the drug along the drug movement path, the current sequence information defining a current applied over time to each of the plurality of electrodes, and outputting the current sequence information to a current source apparatus.

Abstract: A device including processors configured to execute instructions and memories storing the instructions, which when executed by the processors configure the processors to perform an operation for training a transformer model having a plurality of encoders and a plurality of decoders by configuring the processors to identify the batches of training data into a plurality of micro-batches, select layer pairs for the plurality of micro-batches, assemble a processing order of the layer pairs, determining resource information to be allocated to the layer pairs, and allocate resources to the layer pairs based on the determined resource information to be allocated to the layer pairs, dependent con the processing order of the layer pairs.

Abstract: An operating method of an access point (AP) multi-link device (MLD) is provided. The operating method includes: based on at least one non-AP MLD and at least one single link device (SLD) attempting channel access through a first link of the AP MLD, identifying a contention window of the at least one non-AP MLD, based on a ratio of, with respect to the first link, an idle time of the at least one non-AP MLD to an idle time of the at least one SLD; and transmitting a beacon frame including information about the contention window to the at least one non-AP MLD.

Abstract: Disclosed is a three-terminal electro-chemical memory cell with a vertical structure for neuromorphic computation, including a circumferential hole, first and second conductive electrode layers sequentially stacked along an outer surface of the circumferential hole, an electrolyte layer formed along an inner surface of the circumferential hole and connected to one end of each of the first and second conductive electrode layers, and a gate electrode disposed parallel to the electrolyte layer in an inner surface direction of the circumferential hole.

Abstract: A processor-implemented method with checkpointing includes: performing an operation for learning of an artificial neural network (ANN) model; and performing a checkpointing to store information about a state of the ANN model, simultaneously with performing the operation for the learning of the ANN model.

Abstract: A collision safety control system includes a memory storing a collision safety model having a deep learning-based collision safety control logic. The collision safety control system also includes a processor electrically connected to the memory. The processor is configured to, in accordance with the collision safety control logic, train, based on at least one signal including pre-collision data and post-collision data, the collision safety model such that the collision safety model outputs a collision type and a required time-to-fire (RTTF) of passenger protection equipment corresponding to the at least one signal.

Abstract: A display device includes a sensing circuit and a controller which selects a pixel row in a frame period. A vertical blank period of the frame period includes a sensing time in which the sensing circuit performs a sensing operation for the selected pixel row. The sensing circuit measures a first source voltage of a driving transistor of a pixel in the selected pixel row at a first time point of the sensing time, and measures a second source voltage of the driving transistor at a second time point of the sensing time. The controller calculates a threshold voltage parameter and a mobility parameter based on the first and second source voltages, predicts a saturated source voltage of the driving transistor based on the threshold voltage parameter and the mobility parameter, and calculates a threshold voltage of the driving transistor based on the saturated source voltage.

Abstract: An active noise control system selects one reference sensor providing a reference signal with the largest coherence to a noise signal, among a plurality of available reference sensors, as a first entry of a reference sensor set. After selecting the first entry, the active noise control system repeats a process in which a sensor capable of providing the largest information quantity to a current reference sensor set among remaining sensors is selected as a new entry of a reference sensor set, until a desired number of sensors is reached or a desired control level is reached. When the reference sensor set is determined, the active noise control system utilizes the entries of the reference sensor set to generates a noise control signal suitable for canceling the noise signal.

Abstract: A processor-implemented method with homomorphic encryption includes: receiving a first ciphertext corresponding to a first modulus; generating a second ciphertext corresponding to a second modulus by performing modulus raising on the first ciphertext; and performing bootstrapping by encoding the second ciphertext using a commutative property and an associative property of operations included in a rotation operation.

Abstract: Provided are a terminal and a method of accessing an access point (AP) by the terminal, in a wireless communication system. The method includes determining whether to maintain access to a currently accessed radio channel, based on load information of a currently accessed AP, determining a radio channel to be accessed, based on the load information of the AP near the terminal and channel interference information, and accessing the determined radio channel.

Abstract: An electronic apparatus is provided. The electronic device includes: a communication interface comprising communication circuitry; a memory storing information for correcting a signal received from each of a plurality of other electronic devices; and a processor connected to the memory and the communication interface and configured to control the electronic device, wherein the processor is configured to: control the communication interface to perform communication with another electronic device operating as a software enabled access point, identify another electronic device that has transmitted a signal among the plurality of other electronic devices based on the signal based on the signal being received from another electronic device through the communication interface, and correct the received signal based on correction information corresponding to another electronic device among the information stored in the memory.

Abstract: A device and method with continuous real-time autonomous parking planning and control are disclosed. A method of training a path distribution estimation model includes: obtaining initial information including map information, departure information of an autonomous driving device, and destination information of the autonomous driving device; obtaining paths through which the autonomous driving device is movable by inputting the initial information to the path distribution estimation model which predicts the paths; and training the path distribution estimation model to output a path distribution corresponding to the paths based on the initial information.

Abstract: An optical detection system based on image segmentation according an embodiment of the present disclosure includes a pattern generator configured to select at least one region-of-interest from a target image and generate a specific pattern for controlling optical paths for the at least one region-of-interest, a light modulator configured to receive the specific pattern and selectively separate the optical paths for the at least one region-of-interest from the target image, and an optical detection unit configured to detect separated optical signals for the at least one region-of-interest based on different pixels.

Abstract: Described are anti-doppel antibody-drug conjugates, compositions comprising them, and related methods of treating doppel-associated diseases and conditions, including cancer.

Abstract: An error correction code (ECC) decoder includes a syndrome generator and a burst error corrector. The syndrome generator generates global syndrome data and local syndrome data using input data and a parity check matrix based on a constacyclic code. The burst error corrector corrects a correctable error included in the input data using the global syndrome data and the local syndrome data. The input data includes a plurality of data bits arranged along a first direction and a second direction. The ECC decoder simultaneously corrects a single burst error and a multi-bit error. The single burst error occurs on two or more symbols arranged along the first direction in the input data, and each symbol includes two or more data bits. The multi-bit error randomly occurs on two or more data bits in the input data.