Efficient neural-network-based no-reference approach to an overall quality metric for JPEG and JPEG2000 compressed images.
Hantao LiuJudith RediHani AlersRodolfo ZuninoIngrid HeynderickxPublished in: J. Electronic Imaging (2011)
Keyphrases
- quality metrics
- jpeg compressed images
- image quality
- compressed images
- image quality measures
- visual quality
- quality assessment
- quality estimation
- quality evaluation
- image quality assessment
- human visual system
- structural similarity
- peak signal to noise ratio
- compression ratio
- image data
- digital images
- image quality metrics
- image compression
- high quality
- jpeg compression
- dct domain
- compressed domain
- bit rate
- low bit rate
- discrete cosine transform
- reconstructed image
- blocking artifacts
- jpeg images
- pixel values
- data quality
- feature vectors