The challenges includes two tracks.
Track 1 (Multi-modal data): It use the protocol 4 (Cross-ethnicity & PAI) in [1]. Participates can use multiply modalities, such as RGB, depth and IR data. Any extra data or pretrained model cannot be used in track 1.
Track 2 (Single RGB modal data): It use the protocol protocol 4_1 (Cross-ethnicity & PAI) in [1]. But participates are only permitted to use the single RGB modality. Any extra data or pretrained model cannot be used in track 2.