Light field Saliency Dataset (LFSD)




We acquire 100 light fields using the Lytro light field camera. For each light field, we provide: (a) Raw light field data, (b) A rough focal stack  (c) An all-focus image deriving from focal stack  (d) The ground truth corresponding to all-focus image.  To get a valid ground-truth, we ask three individuals to manually segment the saliency regions from the all-focus image. The result are deemed ground truth only when all three results are consistent (i.e., they have an overlap of over 90%).


 [1] Nianyi Li, Jinwei Ye, Yu Ji, Haibin Ling, Jingyi Yu. Saliency Detection on Light Field. CVPR 2014.     [pdf] [Bibtex]

 [2] Nianyi Li, Bilin Sun, Jingyi Yu. A Weighted Sparse Coding Framework for Saliency Detection. CVPR 2015.     [pdf] [Bibtex]


For each example below, we show results by our model and the ground-truth on:




LFSD (100 images)


Code & Results