Apparatus And Method for Detecting An Object Through Analyzing Activation Functions Of Deep Neural Network

摘要

According to an embodiment of the present invention, a method for classifying and detecting a specific object of an image comprises the following steps: classifying a type of specific object of an original image through a convolutional neural network which includes a first convolution layer and a second convolution layer using the output of the first convolution layer; generating an up-sampling image by up-sampling a representative image of a feature map of the second convolution layer based on the size of a feature map of the first convolution layer when the specific object is present; generating a first feature image by performing element-wise multiplication of representative images of a plurality of feature maps of the first convolution layer and the up-sampling image; and generating a detection image based on the first feature image and detecting the shape and position of the specific object from the detection image. Thus, the exact shape and position of an object can be extracted.