Method and device for determining available vehicle boarding area in travel image by using artificial neural network
An embodiment provides an apparatus for determining a vehicle boarding possible area for a driving image using an artificial neural network, including: an image segmentation module that obtains a driving image for a driving direction of a vehicle from a camera module and segments the driving image into a plurality of image strips; a pre-trained boarding availability classification artificial neural network module that uses the image strip as input information and boarding availability information for the image strip as output information; a feature extraction module that extracts an activation map including feature information on the image strip from the boarding availability classification artificial neural network module; and an area information generation module that generates boarding possible area information for the image strip based on the feature information included in the activation map.
1 . An apparatus for determining a vehicle boarding possible area for a driving image using an artificial neural network, comprising:
a processor configured to:
obtain a driving image for a driving direction of a vehicle from a camera,
segment the driving image into a plurality of image strips,
extract an activation map including feature information in the image strip,
generate image segmentation information on the image strip; and
generate boarding possible area information for the image strip based on the feature information included in the activation map and the image segmentation information.
2 . The apparatus for determining the vehicle boarding possible area for the driving image using the artificial neural network of claim 1 , wherein the processor is further configured to generate a corrected activation map by utilizing reliability information for each coordinate included in the image segmentation information based on a representative point of the activation map, and to generate boarding possible area information for the image strip based on the corrected activation map.
3 . The apparatus for determining the vehicle boarding possible area for the driving image using the artificial neural network of claim 1 , wherein
a loss of a segmentation artificial neural network for generating the image segmentation information includes a representative point loss that is a distance difference between representative point information in reference information for the image strip and representative point information in the image segmentation information, and a boundary line loss that is a distance difference between boundary line information in reference information for the image strip and boundary line information in the image segmentation information, and
the segmentation artificial neural network performs training so as to reduce a loss of the segmentation artificial neural network.
4 . The apparatus for determining the vehicle boarding possible area for the driving image using the artificial neural network of claim 1 , wherein the processor is further configured to segment the driving image so that the image strip becomes a rectangular image strip having a vertical length thereof to be longer than a horizontal length thereof.
5 . A method for determining a vehicle boarding possible area for a driving image using an artificial neural network, comprising:
obtaining a driving image for a driving direction of a vehicle from a camera;
segmenting the driving image into a plurality of image strips;
extracting an activation map including feature information in the image strip;
generating image segmentation information on the image strip; and
generating boarding possible area information for the image strip based on the feature information included in the activation map and the image segmentation information.