SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2111.13680 · 2021

GMFlow: Learning Optical Flow via Global Matching

arXiv · PDF · Open in the Atlas

Code that ran

14 of 19 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 6 failures of our instrument; 5 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2111.13680.json · badge: /api/ran/2111.13680.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
haofeixu/gmflow canonical (classifier label)
classifier basis not recorded on the link
11 ran · 4 no recorded run
raymondwang987/nvds community (classifier label)
classifier basis not recorded on the link
3 ran · 1 no recorded run
FunctionStatusWhere it lives
compute_color Ran
ran, honoured the contract we drafted
haofeixu/gmflow/utils/flow_viz.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from tianfr/Semantic-Flow, is pointer only (licence: NOASSERTION) · get_code("abcbf43ad9fe6129")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from tianfr/Semantic-Flow, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

compute_errors Ran
ran on a synthesized input
raymondwang987/nvds/infer_NVDS_dpt_bi.py
code served (permissive licence) · get_code("b6528bfa48fa18a0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Create small tensors for pred, gt, and mask
# pred and gt need to be tensors with .cpu() and .numpy() methods
# mask needs to be a boolean tensor for indexing
pred = torch.randn(6, dtype=torch.float32)
gt = torch.abs(torch.randn(6, dtype=torch.float32)) + 0.1  # ensure positive values to avoid division issues
mask = torch.tensor([True, True, True, False, True, False], dtype=torch.bool)

ARGS = {
    "pred": pred,
    "gt": gt,
    "mask": mask
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_scale_and_shift Ran
ran on a synthesized input
raymondwang987/nvds/infer_NVDS_dpt_bi.py
code served (permissive licence) · get_code("d3e701ed2a0c860a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small tensors for the function
# prediction: shape (batch_size, height, width) - model predictions
# target: shape (batch_size, height, width) - ground truth targets
# mask: shape (batch_size, height, width) - binary mask indicating valid pixels

batch_size = 2
height = 4
width = 3

prediction = torch.randn(batch_size, height, width)
target = torch.randn(batch_size, height, width)
mask = torch.randint(0, 2, (batch_size, height, width)).float()

ARGS = {
    "prediction": prediction,
    "target": target,
    "mask": mask,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
coords_grid Ran
ran, honoured the contract we drafted
haofeixu/gmflow/gmflow/geometry.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from envision-research/driverecon, is pointer only (licence: NONE) · get_code("134980aa63271fbb")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from envision-research/driverecon, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

flow_compute_color Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
haofeixu/gmflow/utils/flow_viz.py
code served (permissive licence) · get_code("215f62561b03c262")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create small test optical flow arrays
# u: horizontal flow component
# v: vertical flow component
u = np.random.randn(4, 3).astype(np.float32) * 5  # 4x3 array of flow values
v = np.random.randn(4, 3).astype(np.float32) * 5  # 4x3 array of flow values

ARGS = {
    "u": u,
    "v": v,
    "convert_to_bgr": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
flow_loss_func Ran
ran; our contract draft was wrong, not the code
haofeixu/gmflow/loss.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from vimstereo/DenVisCoM, is pointer only (licence: NONE) · get_code("715a9e324491547b")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from vimstereo/DenVisCoM, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

flow_to_color Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
haofeixu/gmflow/utils/flow_viz.py
code served (permissive licence) · get_code("d59fe7ab39463182")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family int, shape (4, 3, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small optical flow array with shape [H, W, 2]
# H=4, W=3 to keep it tiny and distinct
height, width = 4, 3
flow_uv = np.random.randn(height, width, 2).astype(np.float32) * 5

ARGS = {
    "flow_uv": flow_uv,
    "clip_flow": 10.0,
    "convert_to_bgr": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 3, 3) — shape (4, 3, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
generate_shift_window_attn_mask Ran
ran on a synthesized input
haofeixu/gmflow/gmflow/transformer.py
code served (permissive licence) · get_code("a69f1d64c1b71d23")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function uses split_feature which is not defined, so we need to create it
def split_feature(x, num_splits, channel_last=False):
    """Mock split_feature function that reshapes the input."""
    if channel_last:
        # x shape: (1, h, w, 1)
        # We need to split along width dimension
        b, h, w, c = x.shape
        # Reshape to split into windows
        x = x.view(b, h, w // num_splits, num_splits, c)
        x = x.permute(0, 1, 3, 2, 4).contiguous()
        x = x.view(-1, h, w // num_splits, c)
    return x

# Inject split_feature into the module namespace for the function to use
import sys
current_module = sys.modules[__name__]
setattr(current_module, 'split_feature', split_feature)

ARGS = {
    'input_resolution': (8, 8),
    'window_size_h': 4,
    'window_size_w': 4,
    'shift_size_h': 2,
    'shift_size_w': 2,
    'device': torch.device('cpu'),
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float — no output shape is stated: observe_signature.py could not observe it (fixture:KeyError)
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
generate_window_grid Ran
ran, honoured the contract we drafted
haofeixu/gmflow/gmflow/geometry.py
code served (permissive licence) · get_code("0823d84f9697f2a7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (4, 3, 2) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

ARGS = {
    "h_min": 0.0,
    "h_max": 10.0,
    "w_min": 0.0,
    "w_max": 20.0,
    "len_h": 4,
    "len_w": 3,
    "device": torch.device("cpu"),
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (4, 3, 2) — shape (4, 3, 2) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
merge_splits Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
haofeixu/gmflow/gmflow/utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from cvlab-kaist/PF3plat, is pointer only (licence: NOASSERTION) · get_code("6dc05be22af3598b")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from cvlab-kaist/PF3plat, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

normalize_img Ran
ran on a synthesized input
haofeixu/gmflow/gmflow/utils.py
code served (permissive licence) · get_code("b1ba789ed7470207")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 5, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create two small image tensors with shape (batch=2, channels=3, height=4, width=4)
# Values in range [0, 255] to match expected input for normalize_img
img0 = torch.randint(0, 256, (2, 3, 4, 4), dtype=torch.float32)
img1 = torch.randint(0, 256, (2, 3, 4, 4), dtype=torch.float32)

ARGS = {
    "img0": img0,
    "img1": img1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 5, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
normalize_prediction_test Ran
ran on a synthesized input
raymondwang987/nvds/smooth_loss.py
code served (permissive licence) · get_code("7e90a3df0831de3d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Construct small target tensor (batch=2, dim=4)
target = torch.randn(2, 4)

# Construct mask tensor with same shape
mask = torch.ones(2, 4, dtype=torch.bool)

ARGS = {
    "target": target,
    "mask": mask,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
single_head_full_attention Ran
ran; our contract draft was wrong, not the code
haofeixu/gmflow/gmflow/transformer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ispc-lab/tma, is pointer only (licence: NONE) · get_code("52b500940a8025a3")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from ispc-lab/tma, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

split_feature Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
haofeixu/gmflow/gmflow/utils.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from cvlab-kaist/PF3plat, is pointer only (licence: NOASSERTION) · get_code("4df5b1dd5d60d6cb")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from cvlab-kaist/PF3plat, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

global_correlation_softmax Not yet run
unverified (harvested, no recorded run)
haofeixu/gmflow/gmflow/matching.py
code served (permissive licence) · get_code("d5a86279192fe3ae")
How it ran

no invocation recorded.

local_correlation_softmax Not yet run
unverified (harvested, no recorded run)
haofeixu/gmflow/gmflow/matching.py
code served (permissive licence) · get_code("d9223ffb6230f989")
How it ran

no invocation recorded.

normalize_coords Not yet run
unverified (harvested, no recorded run)
haofeixu/gmflow/gmflow/geometry.py
code served (permissive licence) · get_code("8d0dd2c61c5c3d89")
How it ran

no invocation recorded.

normalize_prediction_robust Not yet run
unverified (harvested, no recorded run)
raymondwang987/nvds/infer_NVDS_dpt_bi.py
code served (permissive licence) · get_code("959c2141efd7f746")
How it ran

no invocation recorded.

single_head_split_window_attention Not yet run
unverified (harvested, no recorded run)
haofeixu/gmflow/gmflow/transformer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from vimstereo/DenVisCoM, is pointer only (licence: NONE) · get_code("9a948b3b755e6516")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from vimstereo/DenVisCoM, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 19 samples lifted from this paper's linked repositories, 19 carry a function identity and 0 do not. Of the 19 distinct functions with an identity, 16 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 15 later) and 3 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
single_head_full_attention 6 other papers
6 later
2303.11629 (2023-03)
TMA: Temporal Motion Aggregation for Event-based Optical Flow
2403.05327 (2024-03)
DiffSF: Diffusion Models for Scene Flow Estimation
2409.16288 (2024-09)
Self-Supervised Any-Point Tracking by Contrastive Random Walks
and 3 more not named here
compute_color 5 other papers
3 dated earlier, 2 later
2010.14851 (2020-10)
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation
2104.00556 (2021-04)
Deep Two-View Structure-from-Motion Revisited
2104.13918 (2021-04)
High-Resolution Optical Flow from 1D Attention and Correlation
and 2 more not named here
coords_grid 5 other papers
5 later
2312.04524 (2023-12)
RAVE: Randomized Noise Shuffling for Fast and Consistent Video Editing with Diffusion Models
2409.16288 (2024-09)
Self-Supervised Any-Point Tracking by Contrastive Random Walks
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
and 2 more not named here
flow_to_color 5 other papers
4 dated earlier, 1 later
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
2010.14851 (2020-10)
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation
2104.00556 (2021-04)
Deep Two-View Structure-from-Motion Revisited
and 2 more not named here
flow_compute_color 3 other papers
3 dated earlier
2010.14851 (2020-10)
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation
2104.00556 (2021-04)
Deep Two-View Structure-from-Motion Revisited
2104.13918 (2021-04)
High-Resolution Optical Flow from 1D Attention and Correlation
generate_window_grid 3 other papers
3 later
2409.16288 (2024-09)
Self-Supervised Any-Point Tracking by Contrastive Random Walks
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
merge_splits 3 other papers
3 later
2409.16288 (2024-09)
Self-Supervised Any-Point Tracking by Contrastive Random Walks
2410.22128 (2024-10)
PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
single_head_split_window_attention 3 other papers
3 later
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
2602.01724 (2026-02)
DenVisCoM: Dense Vision Correspondence Mamba for Efficient and Real-time Optical Flow and Stereo Estimation
split_feature 3 other papers
3 later
2409.16288 (2024-09)
Self-Supervised Any-Point Tracking by Contrastive Random Walks
2410.22128 (2024-10)
PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
global_correlation_softmax 2 other papers
2 later
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
local_correlation_softmax 2 other papers
2 later
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
normalize_coords 2 other papers
2 later
2410.13862 (2024-10)
DepthSplat: Connecting Gaussian Splatting and Depth
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
flow_loss_func 1 other paper
1 later
2602.01724 (2026-02)
DenVisCoM: Dense Vision Correspondence Mamba for Efficient and Real-time Optical Flow and Stereo Estimation
generate_shift_window_attn_mask 1 other paper
1 later
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
normalize_img 1 other paper
1 later
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion
normalize_prediction_robust 1 other paper
1 later
2503.01257 (2025-03)
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Learning-based optical flow estimation has been dominated with the pipeline of cost volume with convolutions for flow regression, which is inherently limited to local correlations and thus is hard to address the long-standing challenge of large displacements. To alleviate this, the state-of-the-art framework RAFT gradually improves its prediction quality by using a large number of iterative refinements, achieving remarkable performance but introducing linearly increasing inference time. To enable both high accuracy and efficiency, we completely revamp the dominant flow regression pipeline by reformulating optical flow as a global matching problem, which identifies the correspondences by directly comparing feature similarities. Specifically, we propose a GMFlow framework, which consists of three main components: a customized Transformer for feature enhancement, a correlation and softmax layer for global feature matching, and a self-attention layer for flow propagation. We further introduce a refinement step that reuses GMFlow at higher feature resolution for residual flow prediction. Our new framework outperforms 31-refinements RAFT on the challenging Sintel benchmark, while using only one refinement and running faster, suggesting a new paradigm for accurate and efficient optical flow estimation. Code is available at https://github.com/haofeixu/gmflow.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2111.13680")
get_code_for_paper("2111.13680")
have("2111.13680")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2111.13680.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2111.13680.svg)](https://syntology.ai/paper/2111.13680)

Connect an agent — have() is free.