SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.00556 · CVPR · 2021

Deep Two-View Structure-from-Motion Revisited

Kaihao Zhang, Hongdong Li, Yiran Zhong, Yuchao Dai, Jianyuan Wang, Stan Birchfield, Nikolai Smolyanskiy

arXiv · PDF · Open in the Atlas

Code that ran

18 of 21 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; 5 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 12 failures of our instrument; 3 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/2104.00556.json · badge: /api/ran/2104.00556.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
jytime/Deep-SfM-Revisited canonical (classifier label)
classifier basis: census
18 ran · 3 no recorded run
FunctionStatusWhere it lives
compute_color Ran
ran, honoured the contract we drafted
jytime/Deep-SfM-Revisited/flow_viz.py
pointer only (licence: MIT); 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 (MIT), 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_valid_depth_mask Ran
ran, violated the contract we drafted
jytime/Deep-SfM-Revisited/demon_metrics.py
code served (permissive licence) · get_code("0d38c54288d1ecd2")
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 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 numpy as np

# Create valid depth maps with some special values
d1 = np.array([
    [1.0, 2.5, np.inf, 0.0],
    [3.2, np.nan, 1.5, 4.1]
], dtype=np.float32)

d2 = np.array([
    [2.0, 1.5, 5.0, np.inf],
    [0.5, 2.2, np.nan, 3.0]
], dtype=np.float32)

ARGS = {
    "d1": d1,
    "d2": d2
}
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 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
convbn Ran
ran; our contract draft was wrong, not the code
jytime/Deep-SfM-Revisited/models/DICL_shallow.py
code served (permissive licence) · get_code("cd9d415e756ddad5")
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 Sequential. The image it ran in was not recorded. If your copy returns the same type 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 torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "in_planes": 3,
    "out_planes": 64,
    "kernel_size": 3,
    "stride": 1,
    "pad": 1,
    "dilation": 1,
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 and dtype family float describe that coercion, not the output, and are not stated
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
coord2flow Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jytime/Deep-SfM-Revisited/epipolar_utils.py
code served (permissive licence) · get_code("5053cba75a330c97")
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 4, dtype family float, shape (2, 2, 4, 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 torch

torch.manual_seed(0)

# Parameters for coord2flow function
b = 2  # batch size
h = 4  # height
w = 3  # width

# coord1: bx3x(h*w) - batch of 2, 3 channels, 12 spatial locations
coord1 = torch.randn(b, 3, h * w, dtype=torch.float32)

# coord2: bx3x(h*w) - batch of 2, 3 channels, 12 spatial locations
coord2 = torch.randn(b, 3, h * w, dtype=torch.float32)

ARGS = {
    "coord1": coord1,
    "coord2": coord2,
    "b": b,
    "h": h,
    "w": w,
}
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 4, dtype family float, shape (2, 2, 4, 3) — shape (2, 2, 4, 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
disparity_loader_png Ran
ran, honoured the contract we drafted
jytime/Deep-SfM-Revisited/KITTI_loader.py
code served (permissive licence) · get_code("1abf6f49a505fcbe")
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 float, shape (4, 3, 1) 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 os.path
import numpy as np
import tempfile
import cv2

# Create a temporary directory and a test PNG file
temp_dir = tempfile.mkdtemp()
temp_path = os.path.join(temp_dir, "test_disparity.png")

# Create a small test disparity image (uint16 format, as disparity files typically are)
# Create a 4x3 image with some test values
test_disp_data = np.array([[100, 200, 300],
                            [400, 500, 600],
                            [700, 800, 900],
                            [1000, 1100, 1200]], dtype=np.uint16)

# Write the test image using cv2
cv2.imwrite(temp_path, test_disp_data)

# Define the arguments dict
ARGS = {
    "path": temp_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, 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 float, shape (4, 3, 1) — shape (4, 3, 1) 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
flow2coord Ran
ran; our contract draft was wrong, not the code
jytime/Deep-SfM-Revisited/epipolar_utils.py
code served (permissive licence) · get_code("a99f9e38ae607fd8")
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 4, 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 expects flow as input
# flow should be of shape (b, 2, h, w) where:
# b = batch size
# 2 = (dx, dy) optical flow components
# h = height
# w = width

# Create a small optical flow tensor
batch_size = 2
height = 4
width = 3
flow = torch.randn(batch_size, 2, height, width, dtype=torch.float32)

ARGS = {
    "flow": flow
}
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 4, 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
flow_compute_color Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jytime/Deep-SfM-Revisited/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_to_color Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jytime/Deep-SfM-Revisited/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
l1 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jytime/Deep-SfM-Revisited/demon_metrics.py
code served (permissive licence) · get_code("94676aa789a890f5")
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 float32, ndim 0, dtype family float, shape () 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 two small depth maps as numpy arrays
# They must be positive, finite values (no nans, infs, or non-positive)
depth1 = np.array([[1.5, 2.0, 3.5],
                    [2.5, 4.0, 1.2]], dtype=np.float32)

depth2 = np.array([[1.6, 1.9, 3.4],
                    [2.6, 3.9, 1.3]], dtype=np.float32)

ARGS = {
    'depth1': depth1,
    'depth2': depth2
}
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 float32, ndim 0, dtype family float, shape () — shape () 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
l1_inverse Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jytime/Deep-SfM-Revisited/demon_metrics.py
code served (permissive licence) · get_code("8d573359b9ccf5f6")
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 float32, ndim 0, dtype family float, shape () 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 two valid depth maps
# depth1: 4x3 array with positive finite values
depth1 = np.array([
    [1.5, 2.0, 2.5],
    [3.0, 1.2, 4.5],
    [2.8, 3.5, 1.8],
    [2.2, 2.9, 3.1]
], dtype=np.float32)

# depth2: 4x3 array with different positive finite values
depth2 = np.array([
    [1.2, 1.8, 2.3],
    [2.8, 1.5, 4.2],
    [3.0, 3.2, 2.0],
    [2.5, 2.6, 3.3]
], dtype=np.float32)

ARGS = {
    "depth1": depth1,
    "depth2": depth2,
}
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 float32, ndim 0, dtype family float, shape () — shape () 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
load_flow_from_png Ran
ran, honoured the contract we drafted
jytime/Deep-SfM-Revisited/KITTI_loader.py
code served (permissive licence) · get_code("914690083c05e41c")
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 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 numpy as np
import cv2
import tempfile
import os

# Create a temporary PNG file with valid flow data
# The function expects a 3-channel image where channels 2 and 1 contain flow data
# and channel 0 contains validity mask

temp_dir = tempfile.mkdtemp()
png_path = os.path.join(temp_dir, "test_flow.png")

# Create a simple flow image: 4x3 image with 3 channels
# Channel 0: validity (0 = invalid, non-zero = valid)
# Channels 1-2: flow data (will be sliced as [:, :, 2:0:-1])
flow_data = np.zeros((4, 3, 3), dtype=np.uint16)

# Set validity mask in channel 0 (non-zero means valid)
flow_data[:, :, 0] = 1

# Set flow values in channels 1 and 2 (these get reversed and used)
# Using values around 32768 to test the offset operation
flow_data[:, :, 1] = 32768 + 100  # vx component
flow_data[:, :, 2] = 32768 + 200  # vy component

# Create an invalid region (where channel 0 is 0)
flow_data[0, 0, 0] = 0

# Write as PNG
cv2.imwrite(png_path, flow_data)

ARGS = {
    "png_path": png_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, 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 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
load_velodyne_points Ran
ran, honoured the contract we drafted
jytime/Deep-SfM-Revisited/kitti_utils.py
code served (permissive licence) · get_code("8bfc895e86bfc7bc")
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 2, dtype family float, shape (5, 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 numpy as np
import tempfile
import os

# Create a temporary binary file with KITTI point cloud data
# KITTI format: each point is 4 float32 values (x, y, z, reflectance)
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, "points.bin")

# Create sample point cloud data: 5 points with 4 values each
sample_points = np.array([
    [1.0, 2.0, 3.0, 0.5],
    [4.0, 5.0, 6.0, 0.7],
    [7.0, 8.0, 9.0, 0.9],
    [10.0, 11.0, 12.0, 0.3],
    [13.0, 14.0, 15.0, 0.6]
], dtype=np.float32)

sample_points.tofile(temp_file)

ARGS = {
    "filename": temp_file
}
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 2, dtype family float, shape (5, 4) — shape (5, 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
mask_loader_png Ran
ran, honoured the contract we drafted
jytime/Deep-SfM-Revisited/KITTI_loader.py
code served (permissive licence) · get_code("3eee763265fa66e7")
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 float, shape (4, 3, 1) 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 os
import tempfile
import numpy as np
import cv2

# Create a temporary PNG file with a simple mask
temp_dir = tempfile.gettempdir()
temp_path = os.path.join(temp_dir, "test_mask.png")

# Create a simple binary mask image and save it
mask_image = np.zeros((4, 3), dtype=np.uint8)
mask_image[1:3, 1:2] = 255  # Add some non-zero pixels
cv2.imwrite(temp_path, mask_image)

ARGS = {
    "path": temp_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, 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 float, shape (4, 3, 1) — shape (4, 3, 1) 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
predict_flow Ran
ran; our contract draft was wrong, not the code
jytime/Deep-SfM-Revisited/models/DICL_shallow.py
pointer only (licence: MIT); the stored copy get_code serves, from yanqi1811/PWC-Net, is pointer only (licence: NOASSERTION) · get_code("c3f23a008d953af5")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from yanqi1811/PWC-Net, 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.

read_calib_file Ran
ran; our contract draft was wrong, not the code
jytime/Deep-SfM-Revisited/kitti_utils.py
code served (permissive licence) · get_code("ff833c099a80a327")
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 dict. The image it ran in was not recorded. If your copy returns the same type 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
import tempfile
import os

# Create a temporary KITTI calibration file
temp_dir = tempfile.mkdtemp()
calib_file_path = os.path.join(temp_dir, "calib.txt")

# Write a sample KITTI calibration file
calib_content = """P0: 7.070493e+02 0.000000e+00 6.040814e+02 0.000000e+00 0.000000e+00 7.070493e+02 1.805066e+02 -3.797839e+01 0.000000e+00 0.000000e+00 1.000000e+00 5.371000e-03
P1: 7.070493e+02 0.000000e+00 6.040814e+02 4.694443e+01 0.000000e+00 7.070493e+02 1.805066e+02 -3.701007e+00 0.000000e+00 0.000000e+00 1.000000e+00 4.851000e-03
P2: 7.070493e+02 0.000000e+00 6.040814e+02 4.694443e+01 0.000000e+00 7.070493e+02 1.805066e+02 -3.701007e+00 0.000000e+00 0.000000e+00 1.000000e+00 4.851000e-03
P3: 7.070493e+02 0.000000e+00 6.040814e+02 -3.324873e+01 0.000000e+00 7.070493e+02 1.805066e+02 2.372145e+00 0.000000e+00 0.000000e+00 1.000000e+00 2.162000e-03
R0_rect: 9.999239e-01 9.837760e-03 -7.445570e-03 -9.869795e-03 9.999421e-01 -4.278459e-03 7.402527e-03 4.351614e-03 9.999631e-01
Tr_velo_to_cam: 7.533745e-03 -9.999714e-01 -6.166020e-04 -4.069766e-03 1.480249e-02 7.280733e-04 -9.998902e-01 -7.631618e-02 9.998621e-01 7.523790e-03 1.480755e-02 -2.717806e-01
Tr_imu_to_velo: 9.999976e-01 7.553071e-04 -2.035826e-03 -8.086759e-01 -7.854027e-04 9.998898e-01 -1.482298e-02 3.195559e-01 2.024406e-03 1.482454e-02 9.998881e-01 -7.437787e-01
"""

with open(calib_file_path, 'w') as f:
    f.write(calib_content)

ARGS = {
    "path": calib_file_path
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
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
resize_sparse_map Ran
ran; our contract draft was wrong, not the code
jytime/Deep-SfM-Revisited/flow_transforms.py
code served (permissive licence) · get_code("bfabd42bc420f7f0")
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 float, shape (3, 9, 1) 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 sparse flow map
# flow should be a 2D array with shape (height, width)
height, width = 4, 6
flow = np.random.randn(height, width).astype(np.float32)

# valid should be a 2D binary mask with same shape as flow
# indicating which pixels have valid flow
valid = np.random.randint(0, 2, size=(height, width)).astype(np.float32)

# Scaling factors
fx = 1.5
fy = 0.8

ARGS = {
    "flow": flow,
    "valid": valid,
    "fx": fx,
    "fy": fy,
}
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 float, shape (3, 9, 1) — shape (3, 9, 1) 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
sub2ind Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jytime/Deep-SfM-Revisited/kitti_utils.py
code served (permissive licence) · get_code("04e9b96b63844176")
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 int, ndim 0, dtype family int, shape () 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
import numpy as np

torch.manual_seed(0)

# The function sub2ind converts row, col matrix subscripts to linear indices
# It takes:
# - matrixSize: a tuple (m, n) representing matrix dimensions
# - rowSub: row subscript(s)
# - colSub: column subscript(s)

# Let's create valid arguments for a 4x3 matrix
matrixSize = (4, 3)
rowSub = 1
colSub = 2

ARGS = {
    "matrixSize": matrixSize,
    "rowSub": rowSub,
    "colSub": colSub
}
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 int, ndim 0, dtype family int, shape () — shape () 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
weighted_smooth_depth_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jytime/Deep-SfM-Revisited/loss_functions.py
code served (permissive licence) · get_code("dff279c368fc0d20")
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 0, dtype family float, shape () 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)

# Create small tensors with distinct dimensions
# pred_disp: depth prediction map, shape (batch, height, width)
# im: image, shape (batch, channels, height, width)

batch_size = 2
channels = 3
height = 4
width = 5

pred_disp = torch.randn(batch_size, height, width)
im = torch.randn(batch_size, channels, height, width)

ARGS = {
    "pred_disp": pred_disp,
    "im": im
}
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 0, dtype family float, shape () — shape () 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
MSE_loss Not yet run
unverified (harvested, no recorded run)
jytime/Deep-SfM-Revisited/models/CVPMVS.py
code served (permissive licence) · get_code("c904f0e9da049f2a")
How it ran

no invocation recorded.

photometric_reconstruction_loss_Charbonnier Not yet run
unverified (harvested, no recorded run)
jytime/Deep-SfM-Revisited/loss_functions.py
code served (permissive licence) · get_code("48b6598aa92d35a5")
How it ran

no invocation recorded.

sL1_loss Not yet run
unverified (harvested, no recorded run)
jytime/Deep-SfM-Revisited/models/CVPMVS.py
code served (permissive licence) · get_code("646d6725b699091e")
How it ran

no invocation recorded.

Also present in other papers' code

Of 21 samples lifted from this paper's linked repositories, 21 carry a function identity and 0 do not. Of the 21 distinct functions with an identity, 11 are also present in other papers whose code we hold (11 with at least one paper dated earlier, 3 the same month, 7 later) and 10 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
predict_flow 11 other papers
7 dated earlier, 4 later
1709.02371 (2017-09)
PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume
1807.09190 (2018-07)
PReMVOS: Proposal-generation, Refinement and Merging for Video Object Segmentation
1810.10510 (2018-10)
Neighbourhood Consensus Networks
and 8 more not named here
load_velodyne_points 7 other papers
1 dated earlier, 5 later, 1 same year, order not determinable
1806.01260 (2018-06)
Digging Into Self-Supervised Monocular Depth Estimation
Zhou_R-MSFM_Recurrent_Multi-Scale_Feature_Modulation_for_Monocular_Depth_Estimating_ICCV_2021_paper (2021 (year only))
R-MSFM: Recurrent Multi-Scale Feature Modulation for Monocular Depth Estimating
2309.05254 (2023-09)
Towards Better Data Exploitation in Self-Supervised Monocular Depth Estimation
and 4 more not named here
read_calib_file 7 other papers
1 dated earlier, 5 later, 1 same year, order not determinable
1806.01260 (2018-06)
Digging Into Self-Supervised Monocular Depth Estimation
Zhou_R-MSFM_Recurrent_Multi-Scale_Feature_Modulation_for_Monocular_Depth_Estimating_ICCV_2021_paper (2021 (year only))
R-MSFM: Recurrent Multi-Scale Feature Modulation for Monocular Depth Estimating
2309.05254 (2023-09)
Towards Better Data Exploitation in Self-Supervised Monocular Depth Estimation
and 4 more not named here
sub2ind 7 other papers
1 dated earlier, 5 later, 1 same year, order not determinable
1806.01260 (2018-06)
Digging Into Self-Supervised Monocular Depth Estimation
Zhou_R-MSFM_Recurrent_Multi-Scale_Feature_Modulation_for_Monocular_Depth_Estimating_ICCV_2021_paper (2021 (year only))
R-MSFM: Recurrent Multi-Scale Feature Modulation for Monocular Depth Estimating
2309.05254 (2023-09)
Towards Better Data Exploitation in Self-Supervised Monocular Depth Estimation
and 4 more not named here
compute_color 5 other papers
1 dated earlier, 1 the same month, 3 later
2010.14851 (2020-10)
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation
2104.13918 (2021-04)
High-Resolution Optical Flow from 1D Attention and Correlation
2111.13680 (2021-11)
GMFlow: Learning Optical Flow via Global Matching
and 2 more not named here
flow_to_color 5 other papers
2 dated earlier, 1 the same month, 2 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
2104.13918 (2021-04)
High-Resolution Optical Flow from 1D Attention and Correlation
and 2 more not named here
flow_compute_color 3 other papers
1 dated earlier, 1 the same month, 1 later
2010.14851 (2020-10)
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation
2104.13918 (2021-04)
High-Resolution Optical Flow from 1D Attention and Correlation
2111.13680 (2021-11)
GMFlow: Learning Optical Flow via Global Matching
compute_valid_depth_mask 1 other paper
1 dated earlier
1912.09697 (2019-12)
DeepSFM: Structure From Motion Via Deep Bundle Adjustment
l1 1 other paper
1 dated earlier
1912.09697 (2019-12)
DeepSFM: Structure From Motion Via Deep Bundle Adjustment
l1_inverse 1 other paper
1 dated earlier
1912.09697 (2019-12)
DeepSFM: Structure From Motion Via Deep Bundle Adjustment
load_flow_from_png 1 other paper
1 dated earlier
1504.06852 (2015-04)
FlowNet: Learning Optical Flow with Convolutional Networks

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

Two-view structure-from-motion (SfM) is the cornerstone of 3D reconstruction and visual SLAM. Existing deep learning-based approaches formulate the problem by either recovering absolute pose scales from two consecutive frames or predicting a depth map from a single image, both of which are ill-posed problems. In contrast, we propose to revisit the problem of deep two-view SfM by leveraging the well-posedness of the classic pipeline. Our method consists of 1) an optical flow estimation network that predicts dense correspondences between two frames; 2) a normalized pose estimation module that computes relative camera poses from the 2D optical flow correspondences, and 3) a scale-invariant depth estimation network that leverages epipolar geometry to reduce the search space, refine the dense correspondences, and estimate relative depth maps. Extensive experiments show that our method outperforms all state-of-the-art two-view SfM methods by a clear margin on KITTI depth, KITTI VO, MVS, Scenes11, and SUN3D datasets in both relative pose and depth estimation.

For agents

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

get_harvested_code_for_paper("2104.00556")
get_code_for_paper("2104.00556")
have("2104.00556")

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

curl https://syntology.ai/api/ran/2104.00556.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.00556.svg)](https://syntology.ai/paper/2104.00556)

Connect an agent — have() is free.