Image prediction based on neighbor embedding methods

Mehmet Turkan 1 Christine Guillemot 2
1 TEMICS - Digital image processing, modeling and communication
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
2 Sirocco - Analysis representation, compression and communication of visual data
Inria Rennes – Bretagne Atlantique , IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
Abstract : This paper describes two new intraimage prediction methods based on two data dimensionality reduction methods: nonnegative matrix factorization (NMF) and locally linear embedding. These two methods aim at approximating a block to be predicted in the image as a linear combination of -nearest neighbors determined on the known pixels in a causal neighborhood of the input block. Variable can be seen as a parameter controlling some sort of sparsity constraints of the approximation vector. The impact of this parameter as well as of the nonnegativity and sum-to-one constraints for the addressed prediction problem has been analyzed. The prediction and RD performances of these two new image prediction methods have then been evaluated in a complete image coding-and-decoding algorithm. Simulation results show gains up to 2 dB in terms of the PSNR of the reconstructed signal after coding and decoding of the prediction residue when compared with H.264/AVC intraprediction modes, up to 3 dB when compared with template matching, and up to 1 dB when compared with a sparse prediction method.
Document type :
Journal articles
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-00747007
Contributor : Mehmet Turkan <>
Submitted on : Tuesday, October 30, 2012 - 12:07:45 PM
Last modification on : Tuesday, July 23, 2019 - 4:04:10 PM

Identifiers

  • HAL Id : hal-00747007, version 1

Citation

Mehmet Turkan, Christine Guillemot. Image prediction based on neighbor embedding methods. IEEE Transactions on Image Processing, Institute of Electrical and Electronics Engineers, 2012, 21 (4), pp.1885-1898. ⟨hal-00747007⟩

Share

Metrics

Record views

434