    Chinese Handwriting Aesthetic Evaluation Database

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

CHAED includes 1000 images, 10 for each of 100 Chinese characters including 20 characters with single element and 80 with multi components in different structure types including horizontal composition, vertical composition, half surrounding composition and surrounding composition.
We invited users to evaluate each image's visual quality by 3 levels, good, medium and bad through the website http://59.108.48.27/eval.
In "Aesthetic Visual Quality Evaluation of Chinese Handwritings", for every character, the odd number of images are used to train the ANN and the even number of images are used to test it. That's to say, 001.jpg, 003.jpg, 005.jpg, 007.jpg, 009.jpg are the training set and 002.jpg, 004.jpg, 006.jpg, 008.jpg, 010.jpg are the test set.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Characters.txt: The 100 Chinese characters.

EvaluationData.txt: 1000*33 evaluation results, each including the human evaluation score corresponding to a character image.

ScoreAnalysisPerImage.txt: the first 3 columns can be used to find the matched image by charID. Average evaluation score and the evaluation scores' standard deviation are given in the following two columns.

GB*: information about 10 images of one character.
	
	*.jpg: 10 images with minimum bounding box of the black pixels.
	Radical: component information of multi-component character.(If the character has only one component, there should be no Radical folder)
		*_*_cut.bmp: the seperated component images.
        	*_cut.bmp: the left-up point and right-down point of the component's minimum bounding box in the original image. The order of the component is corresponding to the second variable of the component image name.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Please contact us: sunrongju@pku.edu.cn, lianzhouhui@pku.edu.cn