TCNN: A Transformer Convolutional Neural Network for artifact classification in whole slide images.

•Classifying 7 common artifact types and their variations.•Working on 22 different types of tissue with tumoral and non-tumoral parts for the classifying of artifacts.•Robustness to laboratories changes (e.g., light and illumination) because of image processing in HSV colormap.•Robustness against orientation and scaling thanks to Spatial Transformer layers.•Generalization capability on new data of different laboratories.•Flexibility to upgrade to achieve higher accuracy by employing the different versions of EfficientNet CNN as the model backbone.