excitingx / calligraphy-evaluation Goto Github PK

Evaluating Brush Movements for Chinese Calligraphy: A Computer Vision Based Approach

Python 3.6

tensorflow 1.4.1

Keras 2.1.2

sklearn 0.19.1

image size is 1920*1080

git clone --recursive https://github.com/excitingx/Calligraphy-Evaluation

1. preprocessing:

• framing: split video into frames
• image2vec: convert the images to a vector and label it

2. recognization:

MCNN-LSTM_WriteStateReg.py is the main the process of recognization.

3. evaluation:

• similar_feature_data.csv is the similarity data of the writing traces between the strokes of the students and the teacher.
• artificial_score.csv is the scorec by teachers.
• score_regression.py is the regression model used to predict the score.