This paper describes a feature-correction two-stage vector quantization (FC2VQ) algorithm for the compression of photo ID pictures. The FC2VQ method treats different regions in a facial image differently. A region of facial features (ROFF), containing the eyes and the mouth, is detected and rendered more accurately than the rest of the image. The technique can compress a 128 ×128×8-bit (16,384 bytes total) ID image to an average size of 350 bytes. The quality of the compressed images is far superior to that obtained by other methods, including the JPEG standard, at similar compression ratios.
- Facial feature detection
- Facial image compression
- Vector quantization
- Visual communications and image processing
ASJC Scopus subject areas
- Atomic and Molecular Physics, and Optics