Method

NeXtStereo: Directionally Driven Channel Expansion Gives Adaptive Real-Time Stereo [NeXtStereo-M]
[Anonymous Submission]

Submitted on 19 Nov. 2025 23:34 by
[Anonymous Submission]

Running time:0.03 s
Environment:1 core @ 2.5 Ghz (Python)

Method Description:
We present NeXTStereo, a lightweight stereo disparity estimation
network designed for real-time applications. NeXTStereo builds on
Widened ConvNeXtV2 blocks, leveraging the strong generalization
capabilities of the ConvNeXt and ConvNeXtV2 architectures. In
addition, we introduce Directionally Modulated Attention (DMA), a
novel attention mechanism that incorporates geometric priors to
modulate features extracted from a learnable edge extractor.
Unlike existing attention designs that rely on learned edge
masking layer or edges as soft masks, DMA uses directional
geometric cues to improve feature selection, providing stronger
representational capability while maintaining computational
efficiency. This design provides strong representational capacity
while maintaining efficiency compared to similar sized state of the
art models.
Parameters:
8.62M parameters. [4 8 16] numbered WCNXv2 Blocks.
Latex Bibtex:

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 Object Scene Flow for Autonomous Vehicles (CVPR 2015), 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 Object Scene Flow for Autonomous Vehicles (CVPR 2015), 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

Error D1-bg D1-fg D1-all
All / All 1.70 3.03 1.92
All / Est 1.70 3.03 1.92
Noc / All 1.57 2.75 1.77
Noc / Est 1.57 2.75 1.77
This table as LaTeX

Test Image 0

Error D1-bg D1-fg D1-all
All / All 1.68 1.26 1.62
All / Est 1.68 1.26 1.62
Noc / All 1.62 1.26 1.57
Noc / Est 1.62 1.26 1.57
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 1

Error D1-bg D1-fg D1-all
All / All 2.05 2.26 2.08
All / Est 2.05 2.26 2.08
Noc / All 1.98 2.26 2.01
Noc / Est 1.98 2.26 2.01
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 2

Error D1-bg D1-fg D1-all
All / All 2.33 5.03 2.46
All / Est 2.33 5.03 2.46
Noc / All 2.22 5.03 2.36
Noc / Est 2.22 5.03 2.36
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 3

Error D1-bg D1-fg D1-all
All / All 2.84 4.02 2.95
All / Est 2.84 4.02 2.95
Noc / All 2.82 4.02 2.93
Noc / Est 2.82 4.02 2.93
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 4

Error D1-bg D1-fg D1-all
All / All 0.58 1.24 0.69
All / Est 0.58 1.24 0.69
Noc / All 0.55 1.24 0.67
Noc / Est 0.55 1.24 0.67
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 5

Error D1-bg D1-fg D1-all
All / All 2.60 4.08 2.73
All / Est 2.60 4.08 2.73
Noc / All 2.37 4.08 2.53
Noc / Est 2.37 4.08 2.53
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 6

Error D1-bg D1-fg D1-all
All / All 3.01 1.84 2.89
All / Est 3.01 1.84 2.89
Noc / All 3.08 1.84 2.94
Noc / Est 3.08 1.84 2.94
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 7

Error D1-bg D1-fg D1-all
All / All 0.50 4.27 1.24
All / Est 0.50 4.27 1.24
Noc / All 0.51 4.27 1.25
Noc / Est 0.51 4.27 1.25
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 8

Error D1-bg D1-fg D1-all
All / All 0.44 2.11 0.75
All / Est 0.44 2.11 0.75
Noc / All 0.43 2.11 0.74
Noc / Est 0.43 2.11 0.74
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 9

Error D1-bg D1-fg D1-all
All / All 0.48 1.54 0.75
All / Est 0.48 1.54 0.75
Noc / All 0.49 1.61 0.77
Noc / Est 0.49 1.61 0.77
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 10

Error D1-bg D1-fg D1-all
All / All 0.92 1.55 1.06
All / Est 0.92 1.55 1.06
Noc / All 0.93 1.55 1.07
Noc / Est 0.93 1.55 1.07
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 11

Error D1-bg D1-fg D1-all
All / All 0.90 0.41 0.81
All / Est 0.90 0.41 0.81
Noc / All 0.91 0.41 0.82
Noc / Est 0.91 0.41 0.82
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 12

Error D1-bg D1-fg D1-all
All / All 0.69 0.58 0.68
All / Est 0.69 0.58 0.68
Noc / All 0.54 0.58 0.54
Noc / Est 0.54 0.58 0.54
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 13

Error D1-bg D1-fg D1-all
All / All 0.72 0.09 0.64
All / Est 0.72 0.09 0.64
Noc / All 0.62 0.09 0.55
Noc / Est 0.62 0.09 0.55
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 14

Error D1-bg D1-fg D1-all
All / All 1.35 0.00 1.33
All / Est 1.35 0.00 1.33
Noc / All 1.23 0.00 1.21
Noc / Est 1.23 0.00 1.21
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 15

Error D1-bg D1-fg D1-all
All / All 2.89 0.08 2.63
All / Est 2.89 0.08 2.63
Noc / All 2.95 0.08 2.68
Noc / Est 2.95 0.08 2.68
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 16

Error D1-bg D1-fg D1-all
All / All 3.67 0.03 3.13
All / Est 3.67 0.03 3.13
Noc / All 3.49 0.03 2.97
Noc / Est 3.49 0.03 2.97
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 17

Error D1-bg D1-fg D1-all
All / All 0.86 0.12 0.78
All / Est 0.86 0.12 0.78
Noc / All 0.86 0.12 0.78
Noc / Est 0.86 0.12 0.78
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 18

Error D1-bg D1-fg D1-all
All / All 4.99 0.98 3.09
All / Est 4.99 0.98 3.09
Noc / All 4.88 0.98 3.01
Noc / Est 4.88 0.98 3.01
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 19

Error D1-bg D1-fg D1-all
All / All 0.91 0.09 0.82
All / Est 0.91 0.09 0.82
Noc / All 0.92 0.09 0.83
Noc / Est 0.92 0.09 0.83
This table as LaTeX

Input Image

D1 Result

D1 Error




eXTReMe Tracker