Research and Analysis Of Extraction and Recognition INS Criptions of Bronzebased On Improved K-Means and Convo Lutional Neural Network

Abstract

The natural language processing of Bronze inscriptions is the pivotal step to the study of the historical use of Bronze inscriptions, and the recognition of Bronze inscription words relics is the most important part. Due to its long history and complex font, the difficulty of recognition is increased a lot. At present, convolutional neural network has been widely used in the field of photo recognition, but it has few application in the field of Bronze inscriptions recognition. This paper proposes a method of Bronze inscriptions' extract and recognition based on improved k-means and convolutional neural network, using the improved k-means algorithm to extract characters, which makes preparations for the recognition of convolutional neural network. Experiments has shown that the improved method has significantly improved the accuracy and speed of Bronze inscriptions recognition, and it is also considerably helpful to the Bronze inscriptions research.