A Mixed Visual Encoding Model Based on the Larger-Scale Receptive Field for Human Brain Activity