Light field imaging enables us to capture all lightrays in a visual scene. As light fields are four-dimensional,their captures come with an increased amount of informationto take advantage of. This has stimulated ongoing light fieldspecific research into virtual viewpoints and shallow depth offield rendering, commonly called refocusing. However, the com-putation time and memory required to perform these operationscan make tasks such as real-time rendering impractical. Onesolution is to exploit the salient information of light fields tofocus resources on regions that attract visual attention whenusing these algorithms. Although saliency estimation methodsfor light fields have been previously explored, these focus mainlyon salient object segmentation with the goal of generating onesaliency map per light field.Aiming to create a basis for a 4D saliency prediction modelanalogous to light fields, this paper proposes a saliency estimationmethod specific to light fields that considers the refocusingoperation. The proposed method modifies an existing viewrendering algorithm with focus guidance, obtained from the lightfield disparity. This facilitates the construction of saliency mapswithout the need to render the corresponding view itself, whichwill help to speed up processing operations that are compatible.The results show that the proposed saliency estimation approachyields very good predictions of visual attention across multipleplanes of the light field. We anticipate that this approach can beextended for a range of rendering applications.
Due to the page limit of the venue to which we submitted our work, we could not provide all the results in the manuscript. Here we will share additional results for our study.
The data we used were from 4 different databases: Stanford1, EPFL2, HCI Heidelberg3.
|
|
 |
|
|
Boardgames3 |
Region1 |
 |
|
Boardgames3 |
Region2 |
 |
|
Dino3 |
Region1 |
 |
|
Dino3 |
Region2 |
 |
|
Dishes3 |
Region1 |
 |
|
Dishes3 |
Region2 |
 |
|
Friends2 |
Region1 |
 |
|
Friends2 |
Region2 |
 |
|
LegoKnights1 |
Region1 |
 |
|
LegoKnights1 |
Region2 |
 |
|
Medieval3 |
Region1 |
 |
|
Medieval3 |
Region2 |
 |
|
Pens3 |
Region1 |
 |
|
Pens3 |
Region2 |
 |
|
Platonic3 |
Region1 |
 |
|
Platonic3 |
Region2 |
 |
|
Sideboard3 |
Region1 |
 |
|
Sideboard3 |
Region2 |
 |
|
Table3 |
Region1 |
 |
|
Table3 |
Region2 |
 |
|
Tarot-Large1 |
Region1 |
 |
|
Tarot-Large1 |
Region2 |
 |
|
Tarot-Small1 |
Region1 |
 |
|
Tarot-Small1 |
Region2 |
 |
|
Tower3 |
Region1 |
 |
|
Tower3 |
Region2 |
 |
|
Town3 |
Region1 |
 |
|
Town3 |
Region2 |
 |
|
Treasure1 |
Region1 |
 |
|
Treasure1 |
Region2 |
 |
|
Vespa2 |
Region1 |
 |
|
Vespa2 |
Region2 |
 |
|
Vinyl3 |
Region1 |
 |
|
Vinyl3 |
Region2 |
 |