An image processing apparatus comprises a first and a second eyes-and-mouth detecting units, a first and a second skin model generating units, an inside-of-mouth model generating unit, an eyeball model generating unit, a first and a second deformation parameter generating units, and an output unit. The first and second eyes-and-mouth detecting units detect eye boundaries and a mouth boundary from a first and a second face images, respectively. The first and second skin model generating units generate a first and a second skin models, respectively. The inside-of-mouth model generating unit and the eyeball model generating unit generate an inside-of-mouth model and an eyeball model, respectively. The first and second deformation parameter generating units generate a first and a second deformation parameters, respectively. The output unit output the first and second skin models, the first and second deformation parameters, the inside-of-mouth model, and the eyeball model as animation data.