--- dataset_info: features: - name: image dtype: image - name: id dtype: string - name: title dtype: string - name: url dtype: string - name: artistUrl dtype: string - name: artistName dtype: string - name: artistId dtype: string - name: completitionYear dtype: int64 - name: dictionaries sequence: string - name: location dtype: string - name: period struct: - name: id dtype: string - name: title dtype: string - name: serie struct: - name: id dtype: string - name: title dtype: string - name: genres sequence: string - name: styles sequence: string - name: media sequence: string - name: sizeX dtype: float64 - name: sizeY dtype: float64 - name: diameter dtype: float64 - name: galleries sequence: string - name: tags sequence: string - name: description dtype: string - name: width dtype: int64 - name: height dtype: int64 - name: group dtype: string splits: - name: test num_bytes: 304516774.52 num_examples: 6135 download_size: 297553264 dataset_size: 304516774.52 configs: - config_name: default data_files: - split: test path: data/test-* --- # WikiART Face


Figure 1: Pipeline to produce single WikiART Face image.

WikiART Face is a collection of face from paintings. Inspired by the largest human-face image set [CelebA](https://mmlab.ie.cuhk.edu.hk/projects/CelebA.html), we process all the portraits from WikiART with the pipeline described in Figure 1. As the final step, we blur the corner of each image so that model can pay more attention on the face rather than the background. We release both version of the dataset w/wo the final blur step. Below table shows the image size per each art movement. | Art Movement | WikiART Face | |:-----------------------|---------------:| | abstract-expressionism | 19 | | ecole-de-paris | 19 | | expressionism | 435 | | impressionism | 705 | | naive-art-primitivism | 66 | | neo-impressionism | 37 | | neoclassicism | 2 | | post-impressionism | 414 | | realism | 1440 | | rococo | 910 | | romanticism | 1441 | | surrealism | 73 | | symbolism | 574 | | All | 6135 |