WMCA stands for Wide Multi Channel Presentation Attack, and is a dataset for detecting attack presentations in face recognition. Face recognition is vulnerable to attacks such as the use of silicone masks to hinder face recognition. As these attack tools are upgraded, visual spectrum detection alone cannot meet the challenge. Therefore, a face recognition attack and defense dataset based on multi-channel convolutional neural networks was created. The dataset consists of 1941 real short videos from 72 different people. The data was recorded from several channels (color, depth, infrared, and thermal). The dataset was produced within the framework of the Idiap "IARPA BATL" and "H2020 TESLA" projects to study face recognition attack and defense methods.