Human detection in digital videos is challenging since the human appearance may widely vary. Several algorithms to detect humans in digital images have been recently developed, such as the Aggregated Chanel Features (ACF). Most of them are based on features related to the shape. These algorithms give the best results regarding accuracy but generate many false alarms. In this paper, we propose to use motion features in the ACF to accurately detect humans in digital videos. Three motion feature channels are assessed: MBH, IMHcd and WSTD. The IMHcd presented the best results within the ACF. We demonstrate that our proposal returns more accurate results than the original ACF and presents a reduction in false positive detection rate.