PROJECT TITLE :
Building Hierarchical Representations for Oracle Character and Sketch Recognition
During this paper, we have a tendency to study oracle character recognition and general sketch recognition. 1st, a data set of oracle characters, which are the oldest hieroglyphs in China nevertheless stay a half of contemporary Chinese characters, is collected for analysis. Second, typical visual representations in shape- and sketch-related works are evaluated. We tend to analyze the issues suffered when addressing these representations and verify many representation design criteria. Primarily based on the analysis, we propose a novel hierarchical representation that mixes a Gabor-related low-level representation and a sparse-encoder-related mid-level representation. Extensive experiments show the effectiveness of the proposed illustration in each oracle character recognition and general sketch recognition. The proposed illustration is additionally complementary to convolutional neural network (CNN)-based models. We have a tendency to introduce a resolution to mix the proposed illustration with CNN-based mostly models, and achieve better performances over both approaches. This resolution has crushed humans at recognizing general sketches.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here