Abstract: Provided is a method of creating a local database for local optimization of an object detector based on a deep neural network. The method includes performing preprocessing on an image extracted from real-time collected or pre-collected images from an edge device, modeling a static background image based on the image received through the pre-processing unit and calculating a difference image between a current input image and a background model to model a dynamic foreground image, detecting an object image from the image based on a training model, and creating a local database based on the background image, the foreground image synthesized with the background image, and the object image synthesized with the background image.

Abstract: Provided is a method of creating a local database for local optimization of an object detector based on a deep neural network. The method includes performing preprocessing on an image extracted from real-time collected or pre-collected images from an edge device, modeling a static background image based on the image received through the pre-processing unit and calculating a difference image between a current input image and a background model to model a dynamic foreground image, detecting an object image from the image based on a training model, and creating a local database based on the background image, the foreground image synthesized with the background image, and the object image synthesized with the background image.

Abstract: The present invention is directed to solving the existing problems and provides an apparatus and method for optimizing object detection performance by re-learning data specific to an installed location from an online server using a localization module in an edge terminal receiving a fixed image like a closed circuit television (CCTV) camera.

Abstract: Provided are a system and method for preprocessing sequential video images for fire detection based on deep learning and a method of training a deep learning network for fire detection. The system includes a selector configured to select a plurality of sequential images in streaming data as one set, a converter configured to convert color information of the plurality of selected sequential set images into spectral reflectance information using preset standard color space feature information, a filtering part configured to filter the plurality of sequential set images of which the color information is converted into the spectral reflectance information through a fire color filter, and a fire map generator configured to generate one fire map image by compressing the plurality of filtered sequential set images.

Abstract: Provided is a method of separating a foreground and a background by extracting a depth image through a stereo camera, generating an occupancy grid map on the basis of the depth image, predicting a free space, and computing a membership value, the method including setting a threshold value of a foreground object existing in a free space boundary region of the predicted free space, determining whether the membership value reaches the threshold value of the foreground object while the membership value is computed, terminating the computing of the membership value when it is determined that the membership value being computed reaches the threshold value of the foreground object in the determining, and separating a foreground and a background through the computed membership value.

Abstract: Provided is a method of separating a foreground and a background by extracting a depth image through a stereo camera, generating an occupancy grid map on the basis of the depth image, predicting a free space, and computing a membership value, the method including setting a threshold value of a foreground object existing in a free space boundary region of the predicted free space, determining whether the membership value reaches the threshold value of the foreground object while the membership value is computed, terminating the computing of the membership value when it is determined that the membership value being computed reaches the threshold value of the foreground object in the determining, and separating a foreground and a background through the computed membership value.

Abstract: Provided are an obstacle detection apparatus and method using a difference image. The obstacle detection apparatus includes an difference image calculator configured to calculate a difference image from a target image from which an obstacle is to be recognized and a comparison image captured prior to the target image, an obstacle area extractor configured to extract an obstacle area which is predicted as including the obstacle, based on boundary information of the calculated difference image, an obstacle area information calculator configured to analyze image information of the extracted obstacle area to calculate a moving speed and moving direction of the obstacle and calculate a position of the obstacle from a reference positon of a camera, and a dangerous obstacle determiner configured to evaluate a danger grade of the obstacle, based on at least one of the moving speed, moving direction, and position of the obstacle.

Abstract: In a real-time embedded system, if a higher-level interrupt having a higher priority than a lower-level interrupt being processed occurs, the lower-level interrupt is stopped from being processed and the higher-level interrupt is processed. Upon completion of the processing of the higher-level interrupt, delay information about the lower-level interrupt is recorded in a compensation timer register corresponding to the lower-level interrupt, and when the processing is stopped, the lower-level interrupt is restarted. Upon completion of the processing of the lower-level interrupt, the next period of the lower-level interrupt is adjusted based on the delay information recorded in the compensation timer register to compensate for the delay.