A Scene Text Detection Algorithm with Multiple Feature Fusion Based on Multiple Kernel Support Vector Machine

Since the text information in the image includes rich semantic meaning, the detection of scene text has very important meaning. The traditional use of single or certain feature methods to detect text is not ideal in complex scenes, and the use of deep learning methods requires a lot of calculations and a large number of training samples. In the case of a small sample field, the detection effect is not satisfactory. Satisfactory. This paper proposes a method based on the fusion of multiple features such as character geometric features and high-level features, and uses multi-core SVM to detect the detection area. By using multiple feature fusion, the text area can be detected more effectively in complex scenes. Multi-core functions can avoid the limitations of single-core functions, thereby improving its performance. Experimental results show that this method can effectively detect scene text.