The goal of this work is to investigate audiovisual-to-articulatory inversion. It is well established that acoustic-to-articulatory inversion is an under-determined problem. On the other hand, there is strong evidence that human speakers/listeners exploit the multimodality of speech, and more particularly the articulatory cues : the view of visible articulators, i.e. jaw and lips, improves speech intelligibility. It is thus interesting to add constraints provided by the direct visual observation of the speaker's face. Visible data were obtained by stereo-vision and enable the 3D recovery of jaw and lip movements. These data were processed to fit the nature of parameters of Maeda's articulatory model. Inversion experiments show that constraints on visible articulatory parameters enable relevant articulatory trajectories to be recovered and substantially reduce time required to explore the articulatory codebook.