Method

DIFNet: Dual-Information Fusion Network for Depth Completion [mono] [DIFNet]


Submitted on 25 Jul. 2026 05:30 by
Kunyang Wu (JiLin University)

Running time:0.14 s
Environment:GPU @ 2.5 Ghz (Python)

Method Description:
DIFNet is an RGB-guided depth completion network
based on frequency-aware dual-information fusion.
Its Initial Feature Fusion layer performs early
multi-scale interaction between RGB and sparse
depth features. The Dual Stream Modeling block
separates high-frequency structures and low-
frequency regions, processing them with a
spatially aware Mamba architecture and densely
connected convolutions, respectively. A decoder
reconstructs the dense depth map, followed by
confidence-guided spatial propagation for
spatially coherent refinement.
Parameters:
DIFNet-Tiny; RGB + 64-line sparse LiDAR input;
maximum depth 90 m; 352.45 GFLOPs; combined L1/L2
loss with lambda1=lambda2=0.5; batch size 1 for
inference.
Latex Bibtex:
@article{wu2026difnet,
title = {DIFNet: Dual-Information Fusion
Network for Depth Completion},
author = {Wu, Kunyang and Lin, Jun and Miao,
Jiawei and Li, Zhengpeng
and Zhang, Xiucai and Xing, Genyuan
and Fan, Yiyao
and Luo, Jinxin and Zhao, Huanyu and
Liu, Yang
and Zhang, Guanyu},
journal = {Information Fusion},
volume = {125},
pages = {103424},
year = {2026},
doi = {10.1016/j.inffus.2025.103424},
publisher = {Elsevier}
}

Detailed Results

This page provides detailed results for the method(s) selected. For the first 20 test images, the percentage of erroneous pixels is depicted in the table. We use the error metric described in Sparsity Invariant CNNs (THREEDV 2017), which considers a pixel to be correctly estimated if the disparity or flow end-point error is <3px or <5% (for scene flow this criterion needs to be fulfilled for both disparity maps and the flow map). Underneath, the left input image, the estimated results and the error maps are shown (for disp_0/disp_1/flow/scene_flow, respectively). The error map uses the log-color scale described in Sparsity Invariant CNNs (THREEDV 2017), depicting correct estimates (<3px or <5% error) in blue and wrong estimates in red color tones. Dark regions in the error images denote the occluded pixels which fall outside the image boundaries. The false color maps of the results are scaled to the largest ground truth disparity values / flow magnitudes.

Test Set Average

iRMSE iMAE RMSE MAE
Error 1.88 0.83 711.73 193.81
This table as LaTeX

Test Image 0

iRMSE iMAE RMSE MAE
Error 1.72 0.58 791.68 145.19
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 1

iRMSE iMAE RMSE MAE
Error 2.39 0.65 650.46 66.32
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 2

iRMSE iMAE RMSE MAE
Error 1.92 1.34 1079.83 397.84
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 3

iRMSE iMAE RMSE MAE
Error 2.75 1.42 592.08 220.79
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 4

iRMSE iMAE RMSE MAE
Error 2.49 1.31 541.74 205.60
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 5

iRMSE iMAE RMSE MAE
Error 3.54 0.86 929.81 160.27
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 6

iRMSE iMAE RMSE MAE
Error 4.51 0.92 1006.98 159.06
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 7

iRMSE iMAE RMSE MAE
Error 2.86 1.06 639.29 158.44
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 8

iRMSE iMAE RMSE MAE
Error 1.53 0.57 587.42 130.33
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 9

iRMSE iMAE RMSE MAE
Error 1.73 0.98 572.79 180.44
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 10

iRMSE iMAE RMSE MAE
Error 1.80 1.26 777.99 392.04
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 11

iRMSE iMAE RMSE MAE
Error 2.17 1.03 1179.13 369.58
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 12

iRMSE iMAE RMSE MAE
Error 3.77 1.29 1005.59 224.43
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 13

iRMSE iMAE RMSE MAE
Error 1.27 0.74 659.06 184.05
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 14

iRMSE iMAE RMSE MAE
Error 1.16 0.70 555.15 129.78
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 15

iRMSE iMAE RMSE MAE
Error 3.09 1.14 477.99 148.88
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 16

iRMSE iMAE RMSE MAE
Error 1.38 0.63 545.87 156.09
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 17

iRMSE iMAE RMSE MAE
Error 1.24 0.59 551.15 152.54
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 18

iRMSE iMAE RMSE MAE
Error 1.48 0.77 626.86 227.04
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 19

iRMSE iMAE RMSE MAE
Error 1.04 0.74 653.86 206.01
This table as LaTeX

Input Image

D1 Result

D1 Error




eXTReMe Tracker