A special effect image generating apparatus provided with an image conversion processing circuit for posterization or other image conversion, a luminance and chroma condition setting circuit for image extraction by luminance and colors, a mask pattern generator for selection of a mask, and mixer for preparing a video signal, a transformation unit for transformation processing such as movement, and a post video modification unit for addition of a locus etc. The extracted image automatically tracks movement of the image so there is no change in visual effects, images having the same luminance and color conditions can be deleted by mask adjustment, movement of the image can be handled by adjustment of the mask, and highly accurate boundaries can be set by the wave-filtering and shaping circuit. Therefore, a special effect image with distinctive luminance and color can be easily prepared.