CNN LEARNING METHOD AND LEARNING DEVICE FOR OBJECT DETECTOR BASED ON CNN TO BE USED FOR MULTI-CAMERA OR SURROUND VIEW MONITORING USING IMAGE CONCATENATION AND TARGET OBJECT MERGING NETWORK AND TESTING METHOD AND TESTING DEVICE USING THE SAME

摘要

A method is provided to learn the parameters of a CNN-based object detector suitable for customer requirements, such as key performance index using image concatenation and target object integration network. The CNN may be redesigned as the scale of an object changes due to a change in resolution or focal length according to the key performance indicators. The method comprises the steps of: causing a learning apparatus to generate an image processing network, n processed images; Causing an RPN to generate first to nth object proposals respectively in the processed image, and causing an FC layer to generate first to nth object detection information; And allowing the target object integration network to integrate the object proposal and to integrate the object detection information. In this way, the object proposal can be created using Lidar. Through the above method, the accuracy of the 2D bounding box is improved, and can be usefully performed on multiple cameras, surround view monitoring, and the like.