Abstract: A method for switching driving modes of a subject vehicle to support the subject vehicle to perform a platoon driving by using platoon driving information is provided. And the method includes steps of: (a) a basement server, which interworks with the subject vehicle driving in a first mode, acquiring first platoon driving information, to N-th platoon driving information by referring to a real-time platoon driving information DB; (b) the basement server (i) calculating a first platoon driving suitability score to an N-th platoon driving suitability score by referring to first platoon driving parameters to N-th platoon driving parameters and (ii) selecting a target platoon driving group to be including the subject vehicle; (c) the basement server instructing the subject vehicle to drive in a second mode.

Abstract: A method for learning an automatic labeling device for auto-labeling a base image of a base vehicle using sub-images of nearby vehicles is provided. The method includes steps of: a learning device inputting the base image and the sub-images into previous trained dense correspondence networks to generate dense correspondences; and into encoders to output convolution feature maps, inputting the convolution feature maps into decoders to output deconvolution feature maps; with an integer k from 1 to n, generating a k-th adjusted deconvolution feature map by translating coordinates of a (k+1)-th deconvolution feature map using a k-th dense correspondence; generating a concatenated feature map by concatenating the 1-st deconvolution feature map and the adjusted deconvolution feature maps; and inputting the concatenated feature map into a masking layer to output a semantic segmentation image and instructing a 1-st loss layer to calculate 1-st losses and updating decoder weights and encoder weights.

Abstract: A method for enhancing an accuracy of object distance estimation based on a subject camera by performing pitch calibration of the subject camera more precisely with additional information acquired through V2V communication is provided. And the method includes steps of: (a) a computing device, performing (i) a process of instructing an initial pitch calibration module to apply a pitch calculation operation to the reference image, to thereby generate an initial estimated pitch, and (ii) a process of instructing an object detection network to apply a neural network operation to the reference image, to thereby generate reference object detection information; (b) the computing device instructing an adjusting pitch calibration module to (i) select a target object, (ii) calculate an estimated target height of the target object, (iii) calculate an error corresponding to the initial estimated pitch, and (iv) determine an adjusted estimated pitch on the subject camera by using the error.

Abstract: A method for face recognition by using a multiple patch combination based on a deep neural network is provided. The method includes steps of: a face-recognizing device, (a) if a face image with a 1-st size is acquired, inputting the face image into a feature extraction network, to allow the feature extraction network to generate a feature map by applying convolution operation to the face image with the 1-st size, and to generate multiple features by applying sliding-pooling operation to the feature map, wherein the feature extraction network has been learned to extract a feature using a face image for training having a 2-nd size and wherein the 2-nd size is smaller than the 1-st size; and (b) inputting the multiple features into a learned neural aggregation network, to allow the neural aggregation network to aggregate the multiple features and to output an optimal feature for the face recognition.

Abstract: A method for face recognition by using a multiple patch combination based on a deep neural network is provided. The method includes steps of: a face-recognizing device, (a) if a face image with a 1-st size is acquired, inputting the face image into a feature extraction network, to allow the feature extraction network to generate a feature map by applying convolution operation to the face image with the 1-st size, and to generate multiple features by applying sliding-pooling operation to the feature map, wherein the feature extraction network has been learned to extract a feature using a face image for training having a 2-nd size and wherein the 2-nd size is smaller than the 1-st size; and (b) inputting the multiple features into a learned neural aggregation network, to allow the neural aggregation network to aggregate the multiple features and to output an optimal feature for the face recognition.

Abstract: A learning method for generating integrated object detection information by integrating first object detection information and second object detection information is provided. And the method includes steps of: (a) a learning device instructing a concatenating network to generate one or more pair feature vectors; (b) the learning device instructing a determining network to apply FC operations to the pair feature vectors, to thereby generate (i) determination vectors and (ii) box regression vectors; (c) the learning device instructing a loss unit to generate an integrated loss by referring to the determination vectors, the box regression vectors and their corresponding GTs, and performing backpropagation processes by using the integrated loss, to thereby learn at least part of parameters included in the DNN.

Abstract: A learning method for selecting specific information, to be used for updating an HD Map is provided. And the method includes steps of: (a) a learning device instructing a coordinate neural network to generate a local feature map and a global feature vector by applying a coordinate neural network operation to a coordinate matrix; (b) the learning device instructing a determination neural network to generate a first estimated suitability score to an N-th estimated suitability score by applying a determination neural network operation to the integrated feature map; (c) the learning device instructing a loss layer to generate a loss by referring to (i) the first estimated suitability score to the N-th estimated suitability score and (ii) a first Ground Truth(GT) suitability score to an N-th GT suitability score, and perform backpropagation by using the loss, to thereby learn parameters of the determination neural network and the coordinate neural network.

Abstract: A method for remotely controlling at least one autonomous vehicle is provided. The method includes steps of: an autonomous driving control device, (a) on condition that the autonomous driving control device detects driving environment by referring to sensor information and allows the autonomous vehicle to travel on an autonomous driving mode or a manual driving mode, determining whether the autonomous driving control device fails to establish a driving plan by using the driving environment and whether the autonomous driving control device fails to change to the manual driving mode by using the driving environment; and (b) if the autonomous driving control device fails to establish the driving plan or fails to change to the manual driving mode, selecting a remote control mode and transmitting request information to a remote control service providing server, to allow a remote driver to control the autonomous vehicle by using a specific remote vehicle.

Abstract: A method for on-device continual learning of a neural network which analyzes input data is provided to be used for smartphones, drones, vessels, or a military purpose. The method includes steps of: a learning device, (a) sampling new data to have a preset first volume, instructing an original data generator network, which has been learned, to repeat outputting synthetic previous data corresponding to a k-dimension random vector and previous data having been used for learning the original data generator network, such that the synthetic previous data has a second volume, and generating a batch for a current-learning; and (b) instructing the neural network to generate output information corresponding to the batch. The method can be performed by generative adversarial networks (GANs), online learning, and the like. Also, the present disclosure has effects of saving resources such as storage, preventing catastrophic forgetting, and securing privacy.

Abstract: A learning method for performing a seamless parameter switch by using a location-specific algorithm selection for an optimized autonomous driving is provided. And the method includes steps of: (a) a learning device instructing a K-th convolutional layer to apply a convolution operation to K-th training images, to thereby generate K-th feature maps; (b) the learning device instructing a K-th output layer to apply a K-th output operation to the K-th feature maps, to thereby generate K-th estimated autonomous driving source information; (c) the learning device instructing a K-th loss layer to generate a K-th loss by using the K-th estimated autonomous driving source information and its corresponding GT, and then to perform backpropagation by using the K-th loss, to thereby learn K-th parameters of the K-th CNN; and (d) the learning device storing the K-th CNN in a database after tagging K-th location information to the K-th CNN.

Abstract: A method for on-device continual learning of a neural network which analyzes input data is provided to be used for smartphones, drones, vessels, or a military purpose. The method includes steps of: a learning device, (a) sampling new data to have a preset first volume, instructing an original data generator network, which has been learned, to repeat outputting synthetic previous data corresponding to a k-dimension random vector and previous data having been used for learning the original data generator network, such that the synthetic previous data has a second volume, and generating a batch for a current-learning; and (b) instructing the neural network to generate output information corresponding to the batch. The method can be performed by generative adversarial networks (GANs), online learning, and the like. Also, the present disclosure has effects of saving resources such as storage, preventing catastrophic forgetting, and securing privacy.

Abstract: A learning method for transforming a virtual video on a virtual world to a more real-looking video is provided. And the method includes steps of: (a) a learning device instructing a generating CNN to apply a convolutional operation to an N-th virtual training image, N-th meta data and (N-K)-th reference information to generate an N-th feature map; (b) the learning device instructing the generating CNN to apply a deconvolutional operation to the N-th feature map to generate an N-th transformed image; (c) the learning device instructing a discriminating CNN to apply a discriminating CNN operation to the N-th transformed image to generate a category score vector; (d) the learning device instructing the generating CNN to generate a generating CNN loss by referring to the category score vector and its corresponding GT, and to perform backpropagation by referring to the generating CNN loss to learn parameters of the generating CNN.

Abstract: A method for correcting an incorrect angle of a camera is provided. And the method includes steps of: (a) a computing device, generating first reference data or second reference data according to circumstance information by referring to a reference image; (b) the computing device generating a first angle error or a second angle error by referring to the first reference data or the second reference data with vehicle coordinate data; and (c) the computing device instructing a physical rotation module to adjust the incorrect angle by referring to the first angle error or the second angle error.

Abstract: A method for achieving better performance in an autonomous driving while saving computing powers, by using confidence scores representing a credibility of an object detection which is generated in parallel with an object detection process is provided. And the method includes steps of: (a) a computing device acquiring at least one circumstance image on surroundings of a subject vehicle, through at least one panorama view sensor installed on the subject vehicle; (b) the computing device instructing a Convolutional Neural Network(CNN) to apply at least one CNN operation to the circumstance image, to thereby generate initial object information and initial confidence information on the circumstance image; and (c) the computing device generating final object information on the circumstance image by referring to the initial object information and the initial confidence information, with a support of an RL agent.

Abstract: A learning method for generating parameters capable of representing a degree of credibility of an object detection during a process of the object detection is provided. And the method includes steps of: (a) a learning device instructing a convolutional layer to generate a convolutional feature map by applying a convolutional operation to a training image; (b) the learning device instructing an anchor layer to generate an RPN confidence map including RPN confidence scores; (c) the learning device instructing an FC layer to generate CNN confidence scores, to thereby generate a CNN confidence map; and (d) the learning device instructing a loss layer to learn parameters in the CNN and the RPN by performing backpropagation using an RPN loss and a CNN loss, generated by referring to the RPN confidence map, the CNN confidence map, an estimated object detection result and a GT object detection result.

Abstract: A method for learning an automatic labeling device for auto-labeling a base image of a base vehicle using sub-images of nearby vehicles is provided. The method includes steps of: a learning device inputting the base image and the sub-images into previous trained dense correspondence networks to generate dense correspondences; and into encoders to output convolution feature maps, inputting the convolution feature maps into decoders to output deconvolution feature maps; with an integer k from 1 to n, generating a k-th adjusted deconvolution feature map by translating coordinates of a (k+1)-th deconvolution feature map using a k-th dense correspondence; generating a concatenated feature map by concatenating the 1-st deconvolution feature map and the adjusted deconvolution feature maps; and inputting the concatenated feature map into a masking layer to output a semantic segmentation image and instructing a 1-st loss layer to calculate 1-st losses and updating decoder weights and encoder weights.

Abstract: A method for warning by detecting an abnormal state of a driver of a vehicle based on deep learning is provided. The method includes steps of: a driver state detecting device (a) inputting an interior image of the vehicle into a drowsiness detecting network, to detect a facial part of the driver, detect an eye part from the facial part, detect a blinking state of an eye to determine a drowsiness state, and inputting the interior image into a pose matching network, to detect body keypoints of the driver, determine whether the body keypoints match one of preset driving postures, to determine the abnormal state; and (b) if the driver is in a hazardous state referring to part of the drowsiness state and the abnormal state, transmitting information on the hazardous state to nearby vehicles over vehicle-to-vehicle communication to allow nearby drivers to perceive the hazardous state.

Abstract: A method for supporting at least one administrator to evaluate detecting processes of object detectors to provide logical grounds of an autonomous driving is provided. And the method includes steps of: (a) a computing device instructing convolutional layers, included in an object detecting CNN which has been trained before, to generate reference convolutional feature maps by applying convolutional operations to reference images inputted thereto, and instructing ROI pooling layers included therein to generate reference ROI-Pooled feature maps by pooling at least part of values corresponding to ROIs on the reference convolutional feature maps; and (b) the computing device instructing a representative selection unit to classify the reference ROI-Pooled feature maps by referring to information on classes of objects included in their corresponding ROIs on the reference images, and to generate at least one representative feature map per each class.

Abstract: A method for optimizing a hyperparameter of an auto-labeling device performing auto-labeling and auto-evaluating of a training image to be used for learning a neural network is provided for computation reduction and achieving high precision. The method includes steps of: an optimizing device, (a) instructing the auto-labeling device to generate an original image with its auto label and a validation image with its true and auto label, to assort the original image with its auto label into an easy-original and a difficult-original images, and to assort the validation image with its own true and auto labels into an easy-validation and a difficult-validation images; and (b) calculating a current reliability of the auto-labeling device, generating a sample hyperparameter set, calculating a sample reliability of the auto-labeling device, and optimizing the preset hyperparameter set. This method can be performed by a reinforcement learning with policy gradient algorithms.

Abstract: A method for monitoring blind spots of a cycle using a smart helmet for a rider is provided. The method includes steps of: a blind-spot monitoring device, (a) if a video image of 1-st blind spots corresponding to the smart helmet is acquired, instructing an object detector to detect objects on the video image and confirming 1-st objects in the 1-st blind spots; and (b) determining a smart helmet orientation and a cycle traveling direction by referring to sensor information from part of a GPS sensor, an acceleration sensor, and a geomagnetic sensor on the smart helmet, confirming 2-nd objects, among the 1-st objects, in 2-nd blind spots corresponding to the cycle by referring to the smart helmet orientation and the cycle traveling direction, and displaying the 2-nd objects via an HUD or sounding an alarm that the 2-nd objects are in the 2-nd blind spots via a speaker.