Localization of JPEG double compression through multi-domain convolutional neural networks.
Irene AmeriniTiberio UricchioLamberto BallanRoberto CaldelliPublished in: CoRR (2017)
Keyphrases
- multi domain
- convolutional neural networks
- compression algorithm
- image compression
- compression ratio
- lossy compression
- expert group
- compression standards
- trellis coded quantization
- compression rate
- lossless compression
- jpeg xr
- compression scheme
- cross domain
- jpeg images
- transform coding
- convolutional network
- data compression
- domain specific
- arithmetic coding
- progressive transmission
- compressed images
- image coding
- search computing
- lossless coding
- image quality
- arithmetic coder
- compressed domain
- role based access control
- bit rate
- wavelet transform
- bitstream
- discrete cosine transform
- coding method
- jpeg compression
- heterogeneous networks
- visual quality
- coding scheme
- unsupervised learning
- general purpose
- real world