Mid Sweden University

miun.sePublikasjoner
Endre søk
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf
Shearlet Transform Based Prediction Scheme for Light Field Compression
Mittuniversitetet, Fakulteten för naturvetenskap, teknik och medier, Avdelningen för informationssystem och -teknologi. (Realistic3D)
Tampere University of Technology, Finland.
Mittuniversitetet, Fakulteten för naturvetenskap, teknik och medier, Avdelningen för informationssystem och -teknologi.
Tampere University of Technology, Finland.
Vise andre og tillknytning
2018 (engelsk)Konferansepaper, Poster (with or without abstract) (Fagfellevurdert)
Abstract [en]

Light field acquisition technologies capture angular and spatial information ofthe scene. The spatial and angular information enables various post processingapplications, e.g. 3D scene reconstruction, refocusing, synthetic aperture etc at theexpense of an increased data size. In this paper, we present a novel prediction tool forcompression of light field data acquired with multiple camera system. The captured lightfield (LF) can be described using two plane parametrization as, L(u, v, s, t), where (u, v)represents each view image plane coordinates and (s, t) represents the coordinates of thecapturing plane. In the proposed scheme, the captured LF is uniformly decimated by afactor d in both directions (in s and t coordinates), resulting in a sparse set of views alsoreferred to as key views. The key views are converted into a pseudo video sequence andcompressed using high efficiency video coding (HEVC). The shearlet transform basedreconstruction approach, presented in [1], is used at the decoder side to predict thedecimated views with the help of the key views.Four LF images (Truck, Bunny from Stanford dataset, Set2 and Set9 from High DensityCamera Array dataset) are used in the experiments. Input LF views are converted into apseudo video sequence and compressed with HEVC to serve as anchor. Rate distortionanalysis shows the average PSNR gain of 0.98 dB over the anchor scheme. Moreover, inlow bit-rates, the compression efficiency of the proposed scheme is higher compared tothe anchor and on the other hand the performance of the anchor is better in high bit-rates.Different compression response of the proposed and anchor scheme is a consequence oftheir utilization of input information. In the high bit-rate scenario, high quality residualinformation enables the anchor to achieve efficient compression. On the contrary, theshearlet transform relies on key views to predict the decimated views withoutincorporating residual information. Hence, it has inherit reconstruction error. In the lowbit-rate scenario, the bit budget of the proposed compression scheme allows the encoderto achieve high quality for the key views. The HEVC anchor scheme distributes the samebit budget among all the input LF views that results in degradation of the overall visualquality. The sensitivity of human vision system toward compression artifacts in low-bitratecases favours the proposed compression scheme over the anchor scheme.

sted, utgiver, år, opplag, sider
2018.
HSV kategori
Identifikatorer
URN: urn:nbn:se:miun:diva-33356OAI: oai:DiVA.org:miun-33356DiVA, id: diva2:1193525
Konferanse
Data Compression Conference (DCC 2018),Snowbird, Utah, US, March 27 - March 30, 2018
Tilgjengelig fra: 2018-03-27 Laget: 2018-03-27 Sist oppdatert: 2018-06-11

Open Access i DiVA

fulltext(566 kB)944 nedlastinger
Filinformasjon
Fil FULLTEXT01.pdfFilstørrelse 566 kBChecksum SHA-512
f0fd41bd6a05943c5f0f156431b17cb37f95402ae3f021dd6da74e28243c42415f34842c446f44a78098149dca01676a2eebc59cf12338a6dc31dbf51a0449ae
Type fulltextMimetype application/pdf
fulltext(178 kB)128 nedlastinger
Filinformasjon
Fil FULLTEXT02.pdfFilstørrelse 178 kBChecksum SHA-512
b0d1bdfd5b537aef369aceaa619da053872adb00847e98a5fe11fffa1f0bffd01a21e30c55ea96a5b595553a5d3c2144fc3081516cc56ee5bbbc309f899f57ab
Type fulltextMimetype application/pdf

Person

Ahmad, WaqasSjöström, MårtenOlsson, Roger

Søk i DiVA

Av forfatter/redaktør
Ahmad, WaqasSjöström, MårtenOlsson, Roger
Av organisasjonen

Søk utenfor DiVA

GoogleGoogle Scholar
Totalt: 1072 nedlastinger
Antall nedlastinger er summen av alle nedlastinger av alle fulltekster. Det kan for eksempel være tidligere versjoner som er ikke lenger tilgjengelige

urn-nbn

Altmetric

urn-nbn
Totalt: 877 treff
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf