Abstract: A learning method for calculating collision probability, to be used for determining whether it is appropriate or not to switch driving modes of a vehicle capable of an autonomous driving, by analyzing a recent driving route of a driver is provided. And the method includes steps of: (a) a learning device, on condition that a status vector and a trajectory vector are acquired, performing processes of (i) instructing a status network to generate a status feature map and (ii) instructing a trajectory network to generate a trajectory feature map; (b) the learning device instructing a safety network to calculate a predicted collision probability representing a predicted probability of an accident occurrence; and (c) the learning device instructing a loss layer to generate a loss by referring to the predicted collision probability and a GT collision probability, which have been acquired beforehand, to learn at least part of parameters.

Abstract: A method for auto-labeling images by using a class-agnostic refinement module is provided. The method includes steps of: (a) an auto-labeling device inputting the images into a coverage controlling module, to thereby allow the coverage controlling module to detect objects on the images and thus to output first object detection data including first bounding box data and first class data; (b) the auto-labeling device inputting the images and the first bounding box data into the class-agnostic refinement module, to thereby allow the class-agnostic refinement module to detect the objects on the images and thus to generate second bounding box data, and allowing the class-agnostic refinement module to align the first bounding box data and the second bounding box data to thereby output refined bounding box data; and (c) the auto-labeling device generating second object detection data including the first class data and the refined bounding box data.

Abstract: A method of training an auto labeling device is provided. The method includes steps of: a learning device (a) inputting first images into an FPN to generate first pyramid feature maps, instructing an object detection network (ODN) to generate first bounding boxes, an ROI pooling layer to generate first pooled feature maps, and a deconvolution network to generate first segmentation masks, and training the ODN, the FPN, and the deconvolution network and (b) inputting second images into the FPN to generate second pyramid feature maps, and instructing the ODN to generate second bounding boxes, the ROI pooling layer to generate second pooled feature maps, and the deconvolution network to generate second segmentation masks, and inputting the second pooled feature maps into at least one of first and second classifiers to generate per-pixel class scores and mask uncertainty scores, and training one of the first and the second classifiers.

Abstract: A method for training an auto labeling device performing verification using uncertainty scores of auto-labeled labels is provided. The method includes steps of: a learning device (a) (i) inputting first unlabeled images into a feature pyramid network (FPN) to generate first pyramid feature maps, (ii) allowing an object detection network to generate first bounding boxes, and (iii) training the object detection network and the FPN; (b) (i) allowing the FPN to generate second pyramid feature maps and allowing the object detection network to generate second bounding boxes, (ii) instructing an ROI pooling layer to generate pooled feature maps and inputting the pooled feature maps into at least one of a first classifier to generate first class scores and first box uncertainty scores, and a second classifier to generate second class scores and second box uncertainty scores and (iii) training one of the first classifier and the second classifier.

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 training an auto-labeling device is provided. The method includes: (a) inputting a training image to a feature extraction module to generate a feature, (b) inputting the feature to a fitness estimation module to output a fitness value, inputting the feature to a first classification module to output a first class score and a first uncertainty score, inputting the feature to a second classification module to output a second class score and a second uncertainty score, and then generating a scaled second uncertainty score; and (c) (i) training the first classification module and the feature extraction module by referring to the first class score, (ii) training the second classification module and the feature extraction module by referring to the second class score, (iii) updating a scale parameter by referring to the first uncertainty score and the scaled second uncertainty score, and (iv) training the fitness estimation module.

Abstract: A method for generating training data for a deep learning network is provided.

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 planning an autonomous driving by using a V2X communication and an image processing under a road circumstance where both vehicles capable of the V2X communication and vehicles incapable of the V2X communication exist is provided. And the method includes steps of: (a) a computing device, corresponding to a subject autonomous vehicle, instructing a planning module to acquire recognition information on surrounding vehicles including (i) first vehicles capable of a V2X communication and (ii) second vehicles incapable of the V2X communication; (b) the computing device instructing the planning module to select an interfering vehicle among the surrounding vehicles; and (c) the computing device instructing the planning module to generate a potential interference prediction model, and to modify current optimized route information in order to evade a potential interfering action, to thereby generate updated optimized route information of the subject autonomous vehicle.

Abstract: A method for delivering a steering intention of an autonomous driving module to a steering apparatus more accurately by using a reference map is provided. And the method includes steps of: (a) a computing device, if a subject intended steering signal inputted by the autonomous driving module at a current timing is acquired, instructing a signal adjustment module to select, by referring to the reference map, specific reference steering guide values corresponding to the subject intended steering signal; (b) the computing device (i) adjusting the subject intended steering signal by referring to the specific reference steering guide values, in order to generate a subject adjusted steering signal, and (ii) transmitting the subject adjusted steering signal to the steering apparatus, to thereby support the steering apparatus to rotate the subject vehicle by a specific steering angle corresponding to the subject intended steering signal.

Abstract: A method for providing an autonomous driving service platform for autonomous vehicles by using a competitive computing and information fusion is provided.

Abstract: A learning method for detecting driving events occurring during driving, to thereby detect driving scenarios including at least part of the driving events is provided. The method includes: (a) in response to a specific enumerated event vector including each of pieces of information on each of specific driving events as its specific components in a specific order being acquired, a learning device instructing a recurrent neural network (RNN) to apply RNN operations to the specific components of the specific enumerated event vector, to thereby detect a specific predicted driving scenario including the specific driving events; (b) the learning device instructing a loss module to generate an RNN loss by referring to the specific predicted driving scenario and a specific ground-truth (GT) driving scenario and to perform a backpropagation through time (BPTT) by using the RNN loss, to thereby learn at least part of the parameters of the RNN.

Abstract: A learning method for providing a functional safety by warning a driver about a potential dangerous situation by using an explainable AI which verifies detection processes of a neural network for an autonomous driving is provided.

Abstract: A learning method for supporting a safer autonomous driving through a fusion of information acquired from images and communications is provided. And the method includes steps of: (a) a learning device instructing a first neural network and a second neural network to generate an image-based feature map and a communication-based feature map by using a circumstance image and circumstance communication information; (b) the learning device instructing a third neural network to apply a third neural network operation to the image-based feature map and the communication-based feature map to generate an integrated feature map; (c) the learning device instructing a fourth neural network to apply a fourth neural network operation to the integrated feature map to generate estimated surrounding motion information; and (d) the learning device instructing a first loss layer to train parameters of the first to the fourth neural networks.

Abstract: A method for adjusting a position of a driver assistance device according to a driver state is provided. The method includes steps of: a position adjusting device, (a) inputting an upper and a lower body images of a driver, acquired after the driver sits and starts a vehicle, into a pose estimation network, to acquire body keypoints, calculate body part lengths, and adjust a driver's seat position; and (b) while the vehicle is traveling, inputting the upper body image into a face detector to detect a facial part, inputting the facial part into an eye detector to detect an eye part, and inputting the adjusted driver's seat position and 2D coordinates of an eye into a 3D coordinates transforming device, to generate 3D coordinates of the eye referring to the 2D coordinates and the driver's seat position, and adjust a mirror position of the vehicle referring to the 3D coordinates.

Abstract: A method for providing a dynamic adaptive deep learning model other than a fixed deep learning model, to thereby support at least one specific autonomous vehicle to perform a proper autonomous driving according to surrounding circumstances is provided. And the method includes steps of: (a) a managing device which interworks with autonomous vehicles instructing a fine-tuning system to acquire a specific deep learning model to be updated; (b) the managing device inputting video data and its corresponding labeled data to the fine-tuning system as training data, to thereby update the specific deep learning model; and (c) the managing device instructing an automatic updating system to transmit the updated specific deep learning model to the specific autonomous vehicle, to thereby support the specific autonomous vehicle to perform the autonomous driving by using the updated specific deep learning model other than a legacy deep learning model.

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 method for calculating exact location of a subject vehicle by using information on relative distances is provided. And the method includes steps of: (a) a computing device, if a reference image is acquired through a camera on the subject vehicle, detecting reference objects in the reference image; (b) the computing device calculating image-based reference distances between the reference objects and the subject vehicle, by referring to information on reference bounding boxes, corresponding to the reference objects, on the reference image; (c) the computing device (i) generating a distance error value by referring to the image-based reference distances and coordinate-based reference distances, and (ii) calibrating subject location information of the subject vehicle by referring to the distance error value.

Abstract: A method for learning an automatic parking device of a vehicle for detecting an available parking area is provided. The method includes steps of: a learning device, (a) if a parking lot image of an area nearby the vehicle is acquired, (i) inputting the parking lot image into a segmentation network to output a convolution feature map via an encoder, output a deconvolution feature map by deconvoluting the convolution feature map via a decoder, and output segmentation information by masking the deconvolution feature map via a masking layer; (b) inputting the deconvolution feature map into a regressor to generate relative coordinates of vertices of a specific available parking region, and generate regression location information by regressing the relative coordinates; and (c) instructing a loss layer to calculate 1-st losses by referring to the regression location information and an ROI GT, and learning the regressor via backpropagation using the 1-st losses.

Abstract: A method for detecting emergency vehicles in real time, and managing subject vehicles to support the emergency vehicles to drive without interferences from the subject vehicles by referring to detected information on the emergency vehicles is provided. And the method includes steps of: (a) a management server generating metadata on the specific emergency vehicle by referring to emergency circumstance information; (b) the management server generating a circumstance scenario vector by referring to the emergency circumstance information and the metadata, comparing the circumstance scenario vector with reference scenario vectors, to thereby find a specific scenario vector whose similarity score with the circumstance scenario vector is larger than a threshold, and acquiring an emergency reaction command by referring to the specific scenario vector; (c) the management server transmitting the emergency reaction command to each of the subject vehicles.