Abstract: A driver anonymity-ensured shared vehicle-driving information management system includes a main camera, a sensor, an auxiliary camera, a memory unit, and a processor. The processor includes a driving record generation section configured to generate a driving record of the vehicle on the basis of an image taken by the main camera and sensing data detected by the sensor, a facial feature vector extraction section configured to extract a driver facial feature vector by irreversibly encoding the driver image taken by the auxiliary camera so that the driver image is not able to be restored to the original driver image, and an anonymous driver driving information storage section configured to match the driver facial feature vector with the driving record and stores the matched data in the memory unit.

Abstract: The present disclosure relates to a method of selecting an accident image by using speed profile analysis, which can sufficiently secure an available capacity of a storage medium, can reduce the amount of transmission data and a fee therefor, and can prevent a loss of unnecessary management expenses, by selecting an actual accident image by using speed profile analysis before and after the occurrence of an impact event and deleting, from the storage medium, an image having a grade determined to have a low accident possibility or changing a state of the image into an overwriteable state or taking measures for preventing the transmission of the image to a cloud server.

Abstract: A driver anonymity-ensured shared vehicle-driving information management system includes a main camera, a sensor, an auxiliary camera, a memory unit, and a processor. The processor includes a driving record generation section configured to generate a driving record of the vehicle on the basis of an image taken by the main camera and sensing data detected by the sensor, a facial feature vector extraction section configured to extract a driver facial feature vector by irreversibly encoding the driver image taken by the auxiliary camera so that the driver image is not able to be restored to the original driver image, and an anonymous driver driving information storage section configured to match the driver facial feature vector with the driving record and stores the matched data in the memory unit.

Abstract: The present disclosure relates to a method of selecting an accident image using the results of the recognition of an obstacle on a road, which can distinguish between an actual accident image and a fake accident image by previously recognizing an obstacle on a road while a vehicle travels and determining an impact event to have a low accident possibility, the impact event occurring in a section in which the vehicle goes over the obstacle, or suppressing the impact event, and can secure a space of a storage medium by deleting the fake accident image or prevent a data usage fee and unnecessary management expenses by blocking the transmission of the fake accident image to a remote cloud server.

Abstract: Provided is a neural architecture search method based on knowledge distillation, which trains a student network using knowledge acquired from a teacher network and searches a target neural network. The neural architecture search method may include the steps of: (a) extracting an image feature map from a learning model of the student network; (b) calculating a loss function by comparing an image feature map extracted from a learning model of the teacher network to the image feature map extracted in the step (a); (c) selecting a block whose capacity is to be increased and a block whose capacity is to be decreased, for each learning model block of the student network, based on the loss function; and (d) deciding a candidate learning model of the student network according to the block architecture selected in the step (c).

Abstract: The present disclosure relates to a system for predicting a near future location of an object, which predicts a near future location of a dynamic object, through learning for sampling a sample obtained by estimating a location of the dynamic object by learning an image at current timing, which is captured by a camera, based on artificial intelligence and for changing the sample of the dynamic object whose location has been estimated into a Gaussian mixture model (GMM) by using history data.

Abstract: The present disclosure relates to a method of selecting an accident image by using speed profile analysis, which can sufficiently secure an available capacity of a storage medium, can reduce the amount of transmission data and a fee therefor, and can prevent a loss of unnecessary management expenses, by selecting an actual accident image by using speed profile analysis before and after the occurrence of an impact event and deleting, from the storage medium, an image having a grade determined to have a low accident possibility or changing a state of the image into an overwritable state or taking measures for preventing the transmission of the image to a cloud server.

Abstract: The present disclosure relates to a method of selecting an accident image using the results of the recognition of an obstacle on a road, which can distinguish between an actual accident image and a fake accident image by previously recognizing an obstacle on a road while a vehicle travels and determining an impact event to have a low accident possibility, the impact event occurring in a section in which the vehicle goes over the obstacle, or suppressing the impact event, and can secure a space of a storage medium by deleting the fake accident image or prevent a data usage fee and unnecessary management expenses by blocking the transmission of the fake accident image to a remote cloud server.

Abstract: The present disclosure relates to a method of selecting an accident image by using speed profile analysis, which can sufficiently secure an available capacity of a storage medium, can reduce the amount of transmission data and a fee therefor, and can prevent a loss of unnecessary management expenses, by selecting an actual accident image by using speed profile analysis before and after the occurrence of an impact event and deleting, from the storage medium, an image having a grade determined to have a low accident possibility or changing a state of the image into an overwritable state or taking measures for preventing the transmission of the image to a cloud server.

Abstract: The present disclosure relates to a system for predicting a near future location of an object, which predicts a near future location of a dynamic object, through learning for sampling a sample obtained by estimating a location of the dynamic object by learning an image at current timing, which is captured by a camera, based on artificial intelligence and for changing the sample of the dynamic object whose location has been estimated into a Gaussian mixture model (GMM) by using history data.