Abstract: According to an exemplary embodiment of the present disclosure, a method of detecting a change by using a pre-trained artificial neural network model. In particular, according to the present disclosure, a computing device obtains reference image data and comparison target image data corresponding to the reference image data, and detects a change in the comparison target image data relative to the reference image data by using a pre-trained artificial neural network model, and the pre-trained artificial neural network model corresponds to an artificial neural network model pre-trained based on a pair of image data generated based on original image data at a single time point.

Abstract: Disclosed is a computing device for generating weather observation data for solving the problem. The computing device includes: a memory including computer executable components; and a processor executing following computer executable components stored in the memory, and the computer executable components may include an initial ground weather observation data recognition component recognizing observed initial ground weather observation data, and a weather data generation component trained to generate weather data of a gap region on the initial ground weather observation data by using a machine learning module.

Abstract: Disclosed is a method for annotation based on deep learning, which is performed by a computing device. The method may include: performing first learning of an agent model as supervised learning based on a first bounding box of an interest object corresponding to a ground truth (GT) for annotation; and performing second learning of the agent model as reinforcement learning based on a second bounding box of the interest object randomly sampled according to a constraint for geometric transform of a bounding box.

Abstract: Disclosed is a computing device for generating weather observation data for solving the problem. The computing device includes: a memory including computer executable components; and a processor executing following computer executable components stored in the memory, and the computer executable components may include an initial ground weather observation data recognition component recognizing observed initial ground weather observation data, and a weather data generation component trained to generate weather data of a gap region on the initial ground weather observation data by using a machine learning module.

Abstract: Disclosed is a computing device for generating region information of at least one object included in an image, the computing device including: a memory including computer-executable components; and a processor for executing the computer-executable components stored in the memory, in which the computer-executable components include: a key point heat map generation component for generating, for the image, a key point heat map including key point information of the at least one object; a non-rotating bounding box generation component for generating a non-rotating bounding box based on the generated key point heat map; a rotating bounding box generation component for generating a rotating bounding box by using the non-rotating bounding box; and a final bounding box generation component for representing a region occupied by the at least one object in the image by using at least one of the non-rotating bounding box or the rotating bounding box.

Abstract: Disclosed is a method of training an object prediction model by using input data and a discrimination label including a plurality of discrimination information by a computing device including at least one processor which is a training method including: generating a prediction label based on the input data by using the prediction model; generating a loss value based on a discrimination label corresponding to the input data and the prediction label; and training the prediction model based on the loss value.

Abstract: Disclosed is a method for tracking an object, which is performed by a computing device including at least one processor, including: obtaining a query set including one or more query samples from a first frame included in an image sequence including two or more image frames; obtaining a detection set including one or more detection samples from a second frame included in the image sequence; and determining a label corresponding to each query sample included in the query set, based on the label of each detection sample included in the detection set.

Abstract: Disclosed is a method for scheduling of shooting of a satellite image based on deep learning, which is performed by a computing device. The method may include: generating a prediction image and a cloud amount prediction value up to a future time desired by a user based on a pre-shot satellite image by using a pre-trained neural network model; and determining a shooting schedule of a satellite for at least one region of interest based on the cloud amount prediction value.

Abstract: Disclosed is a method for quantifying algal for management of water quality, performed by a computing device. The method may include: receiving a remote sensing image of an object of interest; and predicting a water quality variable based on the remote sensing image using a pre-trained algal estimation model.

Abstract: A method of detecting a target object performed by a computing device including at least one processor according to an exemplary embodiment of the present disclosure may include: receiving an input image; and generating first result information related to an area corresponding to a target object from the input image based on a trained neural network-based detection model.

Abstract: According to an exemplary embodiment of the present disclosure, a method of predicting the amount of precipitation based on deep learning performed by a computing device is disclosed. The method may include: receiving meteorological data measured in a weather observation system; and predicting the amount of precipitation of a region of interest based on the meteorological data by using a deep learning model. In this case, the deep learning model may be pre-trained based on a combination of a first loss function for an error calculation between a prediction value and Ground Truth (GT), and a second loss function for an error calculation different from the first loss function.

Abstract: Disclosed is a method for tracking an object, which is performed by a computing device including at least one processor, including: obtaining a query set including one or more query samples from a first frame included in an image sequence including two or more image frames; obtaining a detection set including one or more detection samples from a second frame included in the image sequence; and determining a label corresponding to each query sample included in the query set, based on the label of each detection sample included in the detection set.

Abstract: A method of detecting a target object performed by a computing device including at least one processor according to an exemplary embodiment of the present disclosure may include: receiving an input image; and generating first result information related to an area corresponding to a target object from the input image based on a trained neural network-based detection model.

Abstract: According to an exemplary embodiment of the present disclosure, a method of classifying a precipitation type based on deep learning performed by a computing device is disclosed. The method may include: receiving first sensor data and second sensor data measured in a satellite; and generating training data based on at least a part of the first sensor data overlapping the second sensor data.

Abstract: Disclosed is a method for processing a radar image performed by a computing device including at least one processor. The method may include: creating a first polarization image by performing a first decomposition operation with respect to an input radar image; creating a synthetic image through an image creation model based on the input radar image; and creating result information through an image processing model based on the first polarization image and the synthetic image.

Abstract: According to an exemplary embodiment of the present disclosure, a method of super resolution imaging based on deep learning performed by a computing device is disclosed. The method may include: receiving a first image and a second image having low resolution compared to the first image; training a first model so as to generate a low-resolution image corresponding to the second image based on the first image; and training a second model so as to generate a high-resolution image corresponding to the first image based on an output image of the first model. In this case, contents of the first image may not correspond to contents of the second image.

Abstract: According to an exemplary embodiment of the present disclosure, disclosed is a method to train a model. The method to train the model may include: computing an input image using a neural network model; computing an accuracy of a prediction box by comparing the prediction box output from the neural network model and a ground truth box; and training the neural network model by performing backpropagation on the neural network model based on the accuracy.

Abstract: According to an exemplary embodiment of the present disclosure, a method of detecting an object is disclosed. The method of detecting an object includes computing an image including an object by using an object detection model including a local block and a non-local block, in which the local block computes a relationship between adjacent pixels included in a feature map, and the non-local block computes a relationship between non-adjacent pixels included in the feature map.

Abstract: Disclosed is a method for image registration performed by a computing device including at least one processor according to some exemplary embodiments of the present disclosure. The method for image registration may include: determining whether to perform preprocessing on a first image and a second image, based on at least one of the number of first pixels of the first image or the number of second pixels of the second image; when performing the preprocessing, generating a first divided image and a second divided image from each of the first image and the second image through a preprocessing process; and registering the first image and the second image, based on the first divided image and the second divided image.