Intelligent target identification method on manifold

• Main Authors: LAN DEYAN, LIU YUNPENG, XIANG WEI, SHI ZELIN, WU SHUHANG, LIU TIANCI
• Format: Patent
• Language: Chinese; English
• Published: 08.06.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses an intelligent target identification method on manifold. The method comprises the steps of original image input, manifold feature modeling, intelligent network learning, network updating optimization and identification result output. The method comprises the following steps: firstly, carrying out manifold feature modeling on an original image by utilizing a covariance feature matrix, and taking the extracted manifold features as the input of an intelligent learning network; then, based on the Riemannian manifold theory and the matrix differential principle, carrying out gradient model derivation on the deep learning network on the manifold, meanwhile, in the model training process, using a back propagation algorithm based on the matrix chain rule for updating the model, converting the weight optimization process into an optimization problem on the Grassmann manifold; and obtaining an optimal parameter value by using a Riemannian optimization method, and finally learning an intelligent