A graph learning approach for light field image compression

Viola, Irene; Maretić, Hermina Petric; Frossard, Pascal; Ebrahimi, Touradj

doi:10.1117/12.2322827

Cited by 7 publications

(3 citation statements)

References 20 publications

(23 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [46], a graph-based transform derived from a coherent super-pixel over-segmentation of the several views is used to encode non-SKVs. In [47], the non-SKVs are encoded using a graph learning approach which estimates the disparity among the views composing the LF. Finally, in [48], the non-SKVs are generated using a shearlet-transform-based prediction scheme which is shown to be efficient when reconstructing densely sampled LFs under low bitrates.…”

Section: Sai-based Related Workmentioning

confidence: 99%

Light Field Image Coding Based on Hybrid Data Representation

et al. 2020

View full text Add to dashboard Cite

This paper proposes a novel efficient light field coding approach based on a hybrid data representation. Current state-of-the-art light field coding solutions either operate on micro-images or sub-aperture images. Consequently, the intrinsic redundancy that exists in light field images is not fully exploited, as is demonstrated. This novel hybrid data representation approach allows to simultaneously exploit four types of redundancies: i) sub-aperture image intra spatial redundancy, ii) sub-aperture image inter-view redundancy, iii) intra-micro-image redundancy, and iv) inter-micro-image redundancy between neighboring micro-images. The proposed light field coding solution allows flexibility for several types of baselines, by adaptively exploiting the most predominant type of redundancy on a coding block basis. To demonstrate the efficiency of using a hybrid representation, this paper proposes a set of efficient pixel prediction methods combined with a pseudo-video sequence coding approach, based on the HEVC standard. Experimental results show consistent average bitrate savings when the proposed codec is compared to relevant state-ofthe-art benchmarks. For lenslet light field content, the proposed coding algorithm outperforms the HEVCbased pseudo-video sequence coding benchmark by an average bitrate savings of 23%. It is shown for the same light field content that the proposed solution outperforms JPEG Pleno verification models MuLE and WaSP, as these codecs are only able to achieve 11% and −14% bitrate savings over the same HEVC-based benchmark, respectively. The performance of the proposed coding approach is also validated for light fields with wider baselines, captured with high-density camera arrays, being able to outperform both the HEVCbased benchmark, as well as MuLE and WaSP.INDEX TERMS Light field representation, light field image coding, HEVC, pseudo-video sequence, spatial pixel prediction, least squares prediction.

show abstract

Section: Sai-based Related Workmentioning

confidence: 99%

Light Field Image Coding Based on Hybrid Data Representation

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Recently several papers [32]- [36] have been published with comparisons to JPEG Pleno VM implementations of WaSP for encoding densely sampled light fields such as the ones obtained with a plenoptic camera. These light fields represent a subset of the JPEG Pleno datasets [19].…”

Section: Lossy Light Field Codingmentioning

confidence: 99%

“…These light fields represent a subset of the JPEG Pleno datasets [19]. Graph learning technique is used in [32] at the encoder to capture the inter-view redundancy of the light field and the resulting graph is transmitted losslessly. A subset of views are encoded using HEVC and the remaining views are reconstructed by solving a minimization problem at the decoder side.…”

Section: Lossy Light Field Codingmentioning

confidence: 99%

Coding of Light Fields Using Disparity-Based Sparse Prediction

Astola

Tăbuş

2019

IEEE Access

View full text Add to dashboard Cite

In this paper we introduce a new light field codec, dubbed WaSPR (warping and sparse prediction on regions), which has additional features and improved performance when compared to our recently introduced WaSP (warping and sparse prediction) codec. We present in detail the overall scheme, including the initial WaSP structure, and the additional new features of WaSPR, including a region based sparse prediction stage, and the addition of a more efficient way to encode the prediction residuals by including separate inter-view coding of residuals obtained at each hierarchical level. The new scheme is demonstrated to improve over both modes of the JPEG Pleno Verification Model software version 2.1, and is shown to provide competitive results when compared against several recently published light field codecs.INDEX TERMS Light field coding, image coding, plenoptic, multi-view.

show abstract

Hybrid Learning-Based And Hevc-Based Coding Of Light Fields

Stepanov

Valenzise

Dufaux

2020

2020 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

Light fields have additional storage requirements compared to conventional image and video signals, and demand therefore an efficient representation. In order to improve coding efficiency, in this work we propose a hybrid coding scheme which combines a learning-based compression approach with a traditional video coding scheme. Their integration offers great gains at low/mid bitrates thanks to the efficient representation of the learning-based approach and is competitive at high bitrates compared to standard tools thanks to the encoding of the residual signal. The proposed approach achieves on average 38% and 31% BD rate saving compared to HEVC and JPEG Pleno transform-based codec, respectively.

show abstract

A graph learning approach for light field image compression

Cited by 7 publications

References 20 publications

Light Field Image Coding Based on Hybrid Data Representation

Light Field Image Coding Based on Hybrid Data Representation

Coding of Light Fields Using Disparity-Based Sparse Prediction

Hybrid Learning-Based And Hevc-Based Coding Of Light Fields

Contact Info

Product

Resources

About