Abstract: A method using deep learning includes: sampling video to generate a video clip composed sampled frames, and generating a set of differential images between the sampled frames; extracting a first feature map with spatial features and a second feature map with temporal features by applying the sampled frames and the set of differential images to a learned deep learning-based behavior recognition model, and generating a spatio-temporal combined feature map for each sampled frame by combining the first feature map and the second feature map; calculating a bi-directional exponential moving average using the spatio-temporal combined feature map, and then adjusting weights of the spatio-temporal combined feature maps; generating a temporal relationship importance map by applying the video clip and performing late fusion by applying the weight-adjusted spatio-temporal combined feature map and the temporal relationship importance map to the learned deep learning-based behavior recognition model, and classifying the be

Abstract: Disclosed are a high dynamic range imaging method and apparatus. The high dynamic range imaging method includes: (a) acquiring a first low dynamic range (LDR) image, a second LDR image, and a third LDR image having different exposure values; (b) applying the first, second, and third LDR images to a weight map estimation model to extract a weight map, and integrating the first, second, and third LDR images using the extracted weight map to generate an input feature map; and (c) applying the input feature map to a Swin-Fourier convolution network model to perform Fourier transform and convolution on the input feature map, and then connecting a reference feature map to generate a high dynamic range (HDR) image.

Abstract: Disclosed are a high dynamic range imaging method and apparatus. The high dynamic range imaging method includes: (a) acquiring a first low dynamic range (LDR) image, a second LDR image, and a third LDR image having different exposure values; (b) applying the first, second, and third LDR images to a weight map estimation model to extract a weight map, and integrating the first, second, and third LDR images using the extracted weight map to generate an input feature map; and (c) applying the input feature map to a Swin-Fourier convolution network model to perform Fourier transform and convolution on the input feature map, and then connecting a reference feature map to generate a high dynamic range (HDR) image.

Abstract: A computing resource management system using software modularization in a cluster computing environment in which computing devices are connected, including an application process running on each computing device and an algorithm processing process configured to run independently of the application process and perform task processing on the application process, the computing resource management system including: a task managing system configured to receive a task request message from an application process requiring a task from each computing device; a process managing system configured to confirm an algorithm processing process of computing devices connected to the cluster computing environment, and determine whether there is an algorithm processing process in an idle state to which the application process requested for a task will be assigned; and a performed managing system configured to confirm a result of an application process whose task is performed by the algorithm processing process.

Abstract: The inventive concept provides a teaching method for teaching a transfer position of a transfer robot. The teaching method includes: searching for an object on which a target object to be transferred by the transfer robot is placed, based on a 3D position information acquired by a first sensor; and acquiring coordinates of a second direction and coordinates of a third direction of the object based on a data acquired from a second sensor which is a different type from the first sensor.

Abstract: Disclosed are a micelle containing bubbles for drug delivery and a method for manufactuing the same. The micelle formed according to the present invention includes microbubbles fromed by evaporating the liquiefied inert gas due to the irradiation of ultrasonic waves to a solution in which the liquefied inert gas is mixed with a drug for disease treatment, and the micelle is formed as a unit structure in which the microbubbles are contained in a cavity and the cavity is enclosed in a shell, and delivers the drug to the human body. Therefore, the micelle can protect a material of the drug, which is delivered to the human body, as well as deliver a larger amount of the drug than the existing technique, and can be easily applied to the treatment of various diseases.

Abstract: Methods of processing a furnace include the step (I) of supporting the furnace with a support member configured to span across and beyond a footprint of a mounting area. A support surface outside of the footprint is configured to support at least one pair of opposite end portions of the support member such that the furnace is suspended over the mounting area. The method may further include the step (II) of levitating the opposite end portions of the support member over the support surface on a cushion of fluid, and the step (III) of moving the furnace together with the support member relative to the support surface while the opposite end portions of the support member are levitated over the support surface with the cushion of fluid.

Abstract: The present invention relates to a system and a method for supporting robotic hardware design, and the method comprises: loading a template which has information on the mechanical structure, operating mechanism, power transmission and motion data of robot modules; modifying the information for use in a simulator for further designing the robot modules; analyzing the behavior of each robot module according to respective motion data; applying the resulting data to modify each robot module according to simulation API; and controlling the selection/combination of robot parts for simulator assembly according to API and the resulting templates, databases on which robot parts are required and databases for information on selection/combination of robot parts and assembly of each robot module. The above system and method provide an optimal way to design robot hardware and reduce the time required for robot hardware design.

Abstract: A content management method and apparatus for storing and deleting content, backing up content, and restoring the backup content are provided. The method includes receiving the content from a first server and storing the content, and generating and storing content information including information about a location from which the content can be downloaded.

Abstract: The present invention relates to a system and a method for supporting robotic hardware design, and the method comprises: loading a template which has information on the mechanical structure, operating mechanism, power transmission and motion data of robot modules; modifying the information for use in a simulator for further designing the robot modules; analyzing the behavior of each robot module according to respective motion data; applying the resulting data to modify each robot module according to simulation API; and controlling the selection/combination of robot parts for simulator assembly according to API and the resulting templates, databases on which robot parts are required and databases for information on selection/combination of robot parts and assembly of each robot module. The above system and method provide an optimal way to design robot hardware and reduce the time required for robot hardware design.

Abstract: There is provided a wide-band direct conversion transmitting apparatus including: a local oscillation unit generating first, second, and third oscillation signal pairs each including a pair of signals having a phase difference of 90°; an image rejection mixer unit mixing baseband transmission signals including an I signal and a Q signal having a phase difference of 90° with the first oscillation signal pair; a harmonic rejection mixer unit mixing each of the first, second, and third oscillation signal pairs with the baseband transmission signals; and an output signal selecting unit selecting output signals from the image rejection mixer unit or from the harmonic rejection mixer unit.

Abstract: Disclosed herein is a measurement apparatus for improving performances of standard cells in a standard cell library when verifying performance of the standard cell library through a ring oscillator among various test element groups (TEGs). The measurement apparatus includes a ring oscillator block activated in response to an enable signal externally inputted thereto for outputting measurement result values, a decoder for selectively outputting one or more of the measurement result values from the ring oscillator block, and a statistics assistor for receiving output values from the decoder for a predetermined period and outputting a maximum value, a minimum value and an average value of the received values.

Abstract: Disclosed herein is a measurement apparatus for improving performances of standard cells in a standard cell library when verifying performance of the standard cell library through a ring oscillator among various test element groups (TEGs). A built-in circuit is used to measure and verify performance of the standard cell library through a TEG. Therefore, it is possible to effectively improve performances of the standard cells in the standard cell library. Particularly, it is possible to not only remove human errors or internal errors of equipment, but also perform the measurement more readily, rapidly and accurately. Further, it is possible to curtail the use of high-performance equipment or manpower and time required in a measurement process.

Abstract: A frequency divider for a wireless communication system is provided. A frequency divider includes a body bias voltage generator and a divider. The body bias voltage generator generates a body bias voltage including a PMOS body bias voltage and an NMOS body bias voltage whose voltage levels are controlled according to an input signal. The divider includes a plurality of flip-flops whose operation points are determined according to the body bias voltage, and generates an output signal by dividing a frequency of the input signal by N. Each of the flip-flops may include a PMOS logic and an NMOS logic. The PMOS logic may include a plurality of PMOS transistors whose operation points are determined according to the PMOS body bias voltage. The NMOS logic may be connected electrically to the PMOS logic and include a plurality of NMOS transistors whose operation points are determined according to the NMOS body bias voltage.

Abstract: A frequency divider for a wireless communication system is provided. A frequency divider includes a body bias voltage generator and a divider. The body bias voltage generator generates a body bias voltage including a PMOS body bias voltage and an NMOS body bias voltage whose voltage levels are controlled according to an input signal. The divider includes a plurality of flip-flops whose operation points are determined according to the body bias voltage, and generates an output signal by dividing a frequency of the input signal by N. Each of the flip-flops may include a PMOS logic and an NMOS logic. The PMOS logic may include a plurality of PMOS transistors whose operation points are determined according to the PMOS body bias voltage. The NMOS logic may be connected electrically to the PMOS logic and include a plurality of NMOS transistors whose operation points are determined according to the NMOS body bias voltage.

Abstract: Clothes have a capability of preventing sleeping respiratory obstruction and/or decubitus, which is worn by a user. The clothes include a plurality of air chambers provided at the rear of the clothes, the air chambers being capable of being shrunk and expanded. An air supply unit supplies air to each of the air chambers, and a pressure detecting unit measures the pressure of each of the air chambers to produce a measured pressure as an electrical signal. A controller receives the measured pressure signal from the pressure detecting unit to supply the air to the air chambers by using the air supply unit when a snoring or a sleep apnea of the user is detected or a predetermined time elapses in the same posture in order to change a posture of the user.

Abstract: An apparatus for preventing a sleeping respiratory obstruction of a sleeper includes first and second air chambers provided to correspond to both sides of a back of a human body, respectively, so as to be shrunk and expanded, an air supply unit for supplying air to the first and second air chambers so as to expand the first and second air chambers, and an operation selection unit for selecting whether to operate one of the first and second air chambers or to alternately operate the first and second air chambers. A controller controls the air supply unit to control air supply and discharge to and from the first or second air chamber so as to operate one of the first and second air chambers or to alternately operate the first and second air chambers according to the selection result in the operation selection unit.

Abstract: There is provided a wide-band direct conversion transmitting apparatus including: a local oscillation unit generating first, second, and third oscillation signal pairs each including a pair of signals having a phase difference of 90°; an image rejection mixer unit mixing baseband transmission signals including an I signal and a Q signal having a phase difference of 90° with the first oscillation signal pair; a harmonic rejection mixer unit mixing each of the first, second, and third oscillation signal pairs with the baseband transmission signals; and an output signal selecting unit selecting output signals from the image rejection mixer unit or from the harmonic rejection mixer unit.

Abstract: Disclosed herein is a measurement apparatus for improving performances of standard cells in a standard cell library when verifying performance of the standard cell library through a ring oscillator among various test element groups (TEGs). A built-in circuit is used to measure and verify performance of the standard cell library through a TEG. Therefore, it is possible to effectively improve performances of the standard cells in the standard cell library. Particularly, it is possible to not only remove human errors or internal errors of equipment, but also perform the measurement more readily, rapidly and accurately. Further, it is possible to curtail the use of high-performance equipment or manpower and time required in a measurement process.

Abstract: Disclosed is a push-push voltage controlled oscillator which obtains output signals having a frequency two times the fundamental resonance frequency of an LC resonator with differential signals having the same amplitude and opposite phases, creating the advantage of high frequency differential outputs which are obtained using such a voltage controlled oscillator, and thus reducing current consumption.