Table 4 Ablation Analysis: Highlighting the Best Performance in bold.

From: Learning to remove occlusions in light field images using multiscale receptive fields and feature pyramid networks

LF Type

Sparse (syn)

Sparse (real)

Dense (syn)

Name

4-Syn

9-Syn

CD

Single Occ

Double Occ

w/o ResAspp

26.42 / 0.854

27.69 / 0.865

25.01 / 0.864

29.33 / 0.815

27.23 / 0.812

w/o RFB

26.45 / 0.850

27.99 / 0.876

24.85 / 0.863

28.73 / 0.810

29.54 / 0.809

w/o RDB

26.24 / 0.845

26.77 / 0.847

25.05 / 0.867

29.00 / 0.832

27.60 / 0.830

w/o FPN

25.96 / 0.831

26.80 / 0.841

24.41 / 0.832

27.94 / 0.819

26.37 / 0.821

w/o Refinement

26.21 / 0.854

27.06 / 0.857

25.04 / 0.860

28.36 / 0.806

27.07 / 0.802

LF-PyrNet (Ours)

27.41 / 0.872

28.13 / 0.870

25.79 / 0.887

31.96 / 0.826

30.11 / 0.828