This paper proposes an approach to recognise human postures in video sequences, which combines a 2D approach with a 3D human model. The 2D approach consists in projections of moving pixels on the reference axis. The 3D model is a realistic articulated human model which is used to obtain reference postures to compare with test postures. We are interested in a set of specific postures which are representative of typical applications in video interpretation. We give results for recognition of general (e.g. standing) and detailed (e.g standing with one arm up) postures. First results show the effectiveness of our approach for recognition of human posture.