Target Recognition Based on CNN with LeakyReLU and PReLU Activation Functions

The ReLU activation function accelerates the convergence of the training process in the classical framework of deep learning. ReLU causes a large part of the network neurons to die. When a very large gradient flows through a ReLU neuron and updates the parameters, it will not activate any data. This paper proposes target recognition based on CNN with LeakyReLU and PReLU activation functions. According to the advantages of ReLU, LeakyReLU function is used to fix a part of the parameters to cope with the gradient death. PReLU parameters combined with PReLU are trained to construct a new CNN framework. Experimental results show the method is effective and feasibile.