Abstract: Disclosed is a motion amplification device including: an encoder receiving a first frame and a second frame arbitrarily adjacent in an image, and decomposing the first frame into first shape information and first texture information and decomposing the second frame into second shape information and second texture information; a first module generating a third frame in which a motion of an object is amplified based on the first shape information, the second shape information, and the second texture information; a second module analyzing an intensity of the motion based on the first shape information, the second shape information, and the first texture information; and a third module generating amplification image data indicating the intensity of the motion on the third frame.

Abstract: A method of analyzing a motion on the basis of feature tracking is disclosed. The method includes the steps of: capturing image frames; filtering a region of interest (ROI) for the captured image frames; tracking a feature in the captured image frames; removing an extreme value on the basis of an optimum model; and outputting an analysis result. Further disclosed is a system of analyzing a motion on the basis of feature tracking, the system comprising a controller configured to perform each step of the method.

Abstract: The present disclosure provides a text-driven motion recommendation and neural mesh stylization system and a method producing human mesh animation using the same. The system comprises at least one instruction stored in a memory, and a processor that executes the at least one instruction, wherein the at least one instruction, when executed by the processor, causes the processor to find raw action labels matching a query given as a text prompt in a human motion dataset stored in a database, encode the raw action labels and the query for vectorizing the raw action labels and the query, and measure similarity between the raw action labels and the query based on the vectorized vectors.

Abstract: Disclosed herein are a system and method for classifying novel class objects. The method of classifying novel class objects includes (a) constructing a novel classifier considering prior knowledge acquired from a base classifier, and (b) learning a parameterized weight coefficient of a novel classifier model during the training of the novel classifier.

Abstract: An image processing device may perform a method for generating 3D HDR radiance fields, and the method may include: receiving an input of low dynamic range (LDR) images, which are captured at various viewpoints; constructing the 3D high dynamic range (HDR) radiance fields from the LDR images captured at the various viewpoints; generating an HDR image by ray marching on a basis of the 3D HDR radiance fields; generating a tone-mapped LDR image from the HDR image through a tone mapping module; calculating a loss value on a basis of the LDR images captured at the various viewpoints and the tone-mapped LDR image; and modifying parameters of the 3D HDR radiance fields and parameters of the tone-mapping module, so as to maximally minimize the loss value. The LDR images captured at the various viewpoints may be images captured under photometric conditions different from each other.

Abstract: A method of recognizing an object based on mask updates is provided. The method includes generating first result information in which a foreground and background of a predetermined input image have been separated from each other, generating second result information in which the results of the object recognition within the input image and a boundary portion of the object have been matched, calculating an objective function to be applied to each mask head based on the first and second result information and parameter information for the first and second result information, updating all of the mask heads based on the objective function, and performing object recognition based on the updated mask.

Abstract: Provided are a data reduction method and a data processing device. The data processing device includes a memory configured to store target data expressed as a vector matrix and instructions for performing control over data reduction and a processor configured to determine low-rank matrices W1 and W2 from which a target parameter matrix W is constructable. The target parameter matrix W is constructed as the Hadamard product between the low-rank matrices W1 and W2.

Abstract: The present disclosure relates to a method and apparatus for training artificial intelligence based on an episodic memory. According to an embodiment of the present disclosure, a method for training artificial intelligence based on an episodic memory may include: constructing an episodic memory by using a feature vector of a training dataset stored in a full memory; obtaining output data by inputting query data into an artificial intelligence model; deriving a similarity between the output data and a feature vector in the constructed episodic memory; and deriving an episode loss function based on the similarity.

Abstract: Disclosed is an herbicidal composition containing: a pyrimidinedione-based compound as an active ingredient; and an efficacy enhancer. The herbicidal composition can increase biological effect expression characteristics of the pyrimidinedione-based composition as a contact herbicide, improve fast-acting property on monocot weeds as well as broadleaf weeds through enhanced spreading ability and penetrating power, and effectively control, at a low dose, even resistant weeds showing resistance to other herbicides. Furthermore, the herbicidal composition retains thermodynamic stability to increase storage stability, and can exhibit uniform and excellent efficacy at the time of chemical treatment through the uniform dispersion of the active ingredient.

Abstract: A system for controlling exposure of a multi-view camera is disclosed. The system includes a plurality of cameras having different views, a single image-based camera parameter calculator calculating single image-based camera parameters for each of the plurality of cameras; an image intensity ratio calculator calculating an image intensity ratio corresponding to an image intensity difference in an overlapped region between image data obtained from the plurality of cameras; and a multi-image-based camera parameter calculator calculating multi-image-based camera parameters for correcting intensity differences between images of the plurality of cameras on the basis of the single image-based camera parameters and the image intensity ratio.

Abstract: A system for controlling exposure of a multi-view camera is disclosed. The system includes a plurality of cameras having different views, a single image-based camera parameter calculator calculating single image-based camera parameters for each of the plurality of cameras; an image intensity ratio calculator calculating an image intensity ratio corresponding to an image intensity difference in an overlapped region between image data obtained from the plurality of cameras; and a multi-image-based camera parameter calculator calculating multi-image-based camera parameters for correcting intensity differences between images of the plurality of cameras on the basis of the single image-based camera parameters and the image intensity ratio.

Abstract: Described are various technologies that pertain to automatically generating looped videos or cinemagraphs by selecting objects to animate from an input video. In one implementation a group of semantically labeled objects from an input video is received. Candidate objects from the input video that can appear as a moving object in an output cinemagraph or looped video are identified. Candidate video loops are generated using the selected candidate objects. One or more of these candidate video loops are then selected to create a final cinemagraph. The selection of candidate video loops used to create the final cinemagraph can be made by a user or by a predictive model trained to evaluate the subjective attractiveness of the candidate video loops.

Abstract: A cinemagraph is generated that includes one or more video loops. A cinemagraph generator receives an input video, and semantically segments the frames to identify regions that correspond to semantic objects and the semantic object depicted in each identified region. Input time intervals are then computed for the pixels of the frames of the input video. An input time interval for a particular pixel includes a per-pixel loop period and a per-pixel start time of a loop at the particular pixel. In addition, the input time interval of a pixel is based, in part, on one or more semantic terms which keep pixels associated with the same semantic object in the same video loop. A cinemagraph is then created using the input time intervals computed for the pixels of the frames of the input video.

Abstract: Provided are a fire detecting apparatus and a method thereof for detecting a fire, the method includes operations of extracting a feature of at least one object in an input image by using a value of a brightness difference between pixels of the input image or by using an RGB value of the input image; converting the extracted feature of the at least one object into an N dimensional feature; and performing Support Vector Machine (SVM) machine learning on the N dimensional feature of the at least one object.

Abstract: A method and system for detecting motion are provided. The method includes: generating a time domain matrix including vectors corresponding to variation of pixel values as elements of the time domain matrix, of a video image including a plurality of frames; generating a motion matrix from which a low frequency area of the video image is removed by multiplying the time domain matrix by a low rank matrix; and generating a result image including a plurality of frames in which vectors, which are elements of the motion matrix, are included as variation of motion pixel values.

Abstract: Provided are a fire detecting apparatus and a method thereof for detecting a fire, the method includes operations of extracting a feature of at least one object in an input image by using a value of a brightness difference between pixels of the input image or by using an RGB value of the input image; converting the extracted feature of the at least one object into an N dimensional feature; and performing Support Vector Machine (SVM) machine learning on the N dimensional feature of the at least one object.

Abstract: A method and system for detecting motion are provided. The method includes: generating a time domain matrix including vectors corresponding to variation of pixel values as elements of the time domain matrix, of a video image including a plurality of frames; generating a motion matrix from which a low frequency area of the video image is removed by multiplying the time domain matrix by a low rank matrix; and generating a result image including a plurality of frames in which vectors, which are elements of the motion matrix, are included as variation of motion pixel values.

Abstract: Disclosed are uracil compounds represented by Formula 1, a method for preparing the compounds, and a herbicide including the same as an active ingredient: wherein R1, R2, R3, R4, R5, X, Y, Z and W are the same as defined in the detailed description.

Abstract: Disclosed are uracil compounds represented by Formula 1, a method for preparing the compounds, and a herbicide including the same as an active ingredient: wherein R1, R2, R3, R4, R5, X, Y, Z and W are the same as defined in the detailed description.