In this work we present a new approach to deal with problems of Horizontal Stereos- copy. We call Horizontal Stereoscopy the stereoscopy made by changing the traditional paradigm, in which we use a vertical screen to present stereoscopic images, to a new paradigm in which the screen is disposed in a horizontal position where the virtual objects are supported.
Although this changing has already been explored since the decade of 1990 by the Computer Graphics community, we believe that this subject can assume a new dimen- sion if approached by the Computer Vision community, replacing the synthetic images by the generation of stereoscopic pairs made by the processing of images captured by cameras.
In this new approach, the many forms to deal with the appropriate processing that produces the stereoscopic effect define the main issue. We show that this subject is quite rich, being able to produce a new family of theorems and algorithms, in some cases, built by the direct application of well known Computer Vision results, and in other cases demanding new ideas, developed using Computer Vision, Projective Geometry and Optimization.