| 摘要 |
PURPOSE: To compress image data in response to a mode such as an image quality emphasis mode, standard mode, and a motion emphasis mode in addition to a CIF mode and a QCIF mode. CONSTITUTION: A CIF signal or a QCIF signal is given to a discrete cosine transformation section 3, in which the signal is converted into 64 sets of DCT transformation coefficients in the unit of blocks each comprising 8×8 picture elements and a threshold level control section 8 replaces DCT transformation coefficient smaller than 6 different kinds of threshold levels detected by a face area detection section 2 with '0' in the image quality emphasis mode, the standard mode, and the motion emphasis mode. In the image quality emphasis mode, the threshold level is set comparatively smaller and not many DCT transformation coefficients are replaced with '0', in the standard mode, the medium threshold level is set and in the motion emphasis mode, the comparatively higher threshold level is set and lots of DCT transformation coefficients are replaced with '0'. Furthermore, in each mode, the threshold level is set comparatively small in the face area and the threshold level is set comparatively higher in the area other than the face area. |
