Skip to content

Gsoc22 aesthetic detector

##Description This is the implementation of aesthetic detector. The class can load model from an pb file which is a pre-trained model of aesthetic classification with 3 classes and calculate the result of the mode

##Unit test Unit test on 3 good images from AVA dataset, and 3 bad images that are already in the data of unit test of IQS

##TODO The management of model file is not implemented yet, the reason is that pb file of the pretrained model is a large file ( 200MB) to store on gitlab repo of digikam. There should be another solution to store or download this file. Temporary, the model can be downloaded from here

Merge request reports