SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.02431 · CVPR · 2022

Rethinking Efficient Lane Detection via Curve Modeling

Ke Xu, Min Wang, Xin Tan, Lizhuang Ma, Zhengyang Feng, Shaohua Guo

arXiv · PDF · Open in the Atlas

Code that ran

12 of 12 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; 0 honoured the contract we drafted, 0 violated it and 11 have no contract verdict; 1 failure of our instrument; 0 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/2203.02431.json · badge: /api/ran/2203.02431.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
voldemortX/pytorch-auto-drive canonical (classifier label)
classifier basis not recorded on the link
12 ran · 0 no recorded run
FunctionStatusWhere it lives
cmd_dict Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/args.py
code served (permissive licence) · get_code("a130119faaa50abe")
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 torch

torch.manual_seed(0)

def cmd_dict(x):
    options = x.split()
    res = {}
    for o in options:
        kv = o.split('=', maxsplit=1)
        try:
            v = eval(kv[1])
        except:
            v = kv[1]
        res[kv[0]] = v
    return res

x = "learning_rate=0.001 batch_size=32 model_name=ResNet epochs=100"

ARGS = {"x": x}
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 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
count_one_dcn_v2 Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/custom_op_flop_counters.py
code served (permissive licence) · get_code("aedbfbecfdc42cac")
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

# Analyze the function signature and body
# count_one_dcn_v2(fW, fH, C_in, C_out)
# Parameters:
# - fW: feature map width (used in arithmetic, appears to be an int)
# - fH: feature map height (used in arithmetic, appears to be an int)
# - C_in: input channels (used in arithmetic, appears to be an int)
# - C_out: output channels (used in arithmetic, appears to be an int)

# All parameters are used in arithmetic operations to compute MACs
# They should be integer values representing dimensions

torch.manual_seed(0)

ARGS = {
    "fW": 16,
    "fH": 14,
    "C_in": 64,
    "C_out": 128,
}
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
counter_conv Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/custom_op_flop_counters.py
code served (permissive licence) · get_code("7f32bc2caccead61")
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 1, dtype family float, shape (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 torch

torch.manual_seed(0)

ARGS = {
    "bias": 1,
    "kernel_size": 3,
    "output_size": 5,
    "in_channel": 4,
    "group": 2
}
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 1, dtype family float, shape (1,) — shape (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
cubic_bezier_curve_segment Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/curve_utils.py
code served (permissive licence) · get_code("7737677cd5bc1e4a")
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 (2, 4, 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)

# Create control points: shape (B, 4, 2) where B=2, 4 control points, 2D coordinates
control_points = torch.randn(2, 4, 2, dtype=torch.float32)

# Create sample points: shape (B, N, 2) where B=2, N=5 sample points, 2D coordinates
# Values should be in roughly [0, 1] range to pass the valid_points check
sample_points = torch.rand(2, 5, 2, dtype=torch.float32) * 0.8 + 0.1

ARGS = {
    "control_points": control_points,
    "sample_points": sample_points,
}
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 (2, 4, 2) — shape (2, 4, 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
get_lane Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/lane_det_utils.py
code served (permissive licence) · get_code("8b12f4ee1d612dea")
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 1, dtype family float, shape (5,) 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 probability map for a single lane (height=100, width=160)
prob_map = np.random.rand(100, 160).astype(np.float32)
# Add some peak values to make detection realistic
prob_map[30, 80] = 0.95
prob_map[40, 85] = 0.92
prob_map[50, 90] = 0.93

# Parameters
gap = 20  # gap in pixels
ppl = 5   # points per lane
thresh = 0.5  # probability threshold
resize_shape = (200, 320)  # target resize shape (H, W)
dataset = 'culane'  # dataset type

ARGS = {
    'prob_map': prob_map,
    'gap': gap,
    'ppl': ppl,
    'thresh': thresh,
    'resize_shape': resize_shape,
    'dataset': dataset
}
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 1, dtype family float, shape (5,) — shape (5,) 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
get_valid_points Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/curve_utils.py
code served (permissive licence) · get_code("fac83b577eebb6e3")
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 1, dtype family float, shape (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)

# The function expects a tensor 'points' with shape (..., 2)
# It checks if points[..., 0] and points[..., 1] are in range (0, 1)
# Create a small tensor with the required shape
points = torch.tensor([
    [0.5, 0.5],
    [1.5, 0.3],
    [0.2, 0.8],
    [0.9, 0.1]
], dtype=torch.float32)

ARGS = {
    "points": points
}
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 1, dtype family float, shape (4,) — shape (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
lane_as_segmentation_inference Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/lane_det_utils.py
code served (permissive licence) · get_code("b82a202cd3460d27")
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 list, 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
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

class MockLaneNet(nn.Module):
    """Minimal net that produces outputs expected by lane_as_segmentation_inference"""
    def __init__(self):
        super().__init__()
        self.conv = nn.Conv2d(3, 5, kernel_size=3, padding=1)  # 5 classes (including background)
        self.lane_head = nn.Conv2d(3, 4, kernel_size=3, padding=1)  # 4 lanes
    
    def forward(self, x):
        segmentation = self.conv(x)  # (batch, 5, h, w)
        lane_existence = self.lane_head(x)  # (batch, 4, h, w)
        lane_existence = torch.mean(lane_existence, dim=(2, 3), keepdim=True)  # Global average to get lane confidence
        lane_existence = lane_existence.squeeze(-1).squeeze(-1)  # (batch, 4)
        return {
            'out': segmentation,
            'lane': lane_existence
        }

net = MockLaneNet()

# Batch size 2, 3 channels (RGB), height 32, width 24
inputs = torch.randn(2, 3, 32, 24)

# input_sizes: [original_size, target_resize_size]
# Original size during inference
input_sizes = [(32, 24), (64, 48)]

gap = 20
ppl = 3
thresh = 0.3
dataset = 'culane'
max_lane = 0
forward = True

ARGS = {
    'net': net,
    'inputs': inputs,
    'input_sizes': input_sizes,
    'gap': gap,
    'ppl': ppl,
    'thresh': thresh,
    'dataset': dataset,
    'max_lane': max_lane,
    'forward': forward
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family float — a list 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
prob_to_lines Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/lane_det_utils.py
code served (permissive licence) · get_code("e9ee961ae8ccd679")
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 list. 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 cv2

# Create a small segmentation prediction map
# Shape: (num_classes, height, width)
# num_classes = 5 (background + 4 lanes)
# height = 128, width = 256
np.random.seed(0)
seg_pred = np.random.rand(5, 128, 256).astype(np.float32)

# Make some channels have higher probabilities to simulate lane detection
seg_pred[1, 60:80, 50:100] = np.random.rand(20, 50) * 0.5 + 0.5  # Lane 1
seg_pred[2, 60:80, 150:200] = np.random.rand(20, 50) * 0.5 + 0.5  # Lane 2
seg_pred[3, 60:80, 200:250] = np.random.rand(20, 50) * 0.5 + 0.5  # Lane 3

# Existence array for each lane (excluding background)
# Length should be num_classes - 1
exist = [True, True, True, True]

# Resize shape (target output shape)
resize_shape = (256, 512)

# Other parameters
smooth = True
gap = 20
ppl = 10  # points per lane
thresh = 0.3
dataset = 'culane'

ARGS = {
    'seg_pred': seg_pred,
    'exist': exist,
    'resize_shape': resize_shape,
    'smooth': smooth,
    'gap': gap,
    'ppl': ppl,
    'thresh': thresh,
    'dataset': dataset
}
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 list — a list output has no array shape of its own; none is 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
read_config Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/args.py
code served (permissive licence) · get_code("e75a411f8bbc28ca")
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 os
import tempfile
import torch

# Create a temporary Python config file
with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as f:
    config_file = f.name
    f.write("""
# Sample configuration file
MODEL_NAME = "test_model"
LEARNING_RATE = 0.001
BATCH_SIZE = 32
NUM_EPOCHS = 10
USE_CUDA = False
CONFIG_DICT = {"key": "value"}
NESTED_LIST = [1, 2, 3]

def some_function():
    pass

__private_var = "should_not_be_included"
""")

ARGS = {
    "config_path": config_file
}
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 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
reduce_dict Ran
ran; our contract draft was wrong, not the code
voldemortX/pytorch-auto-drive/utils/ddp_utils.py
code served (permissive licence) · get_code("d2092a95e79d2d3b")
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 torch
import torch.distributed as dist

torch.manual_seed(0)

# Create a simple input_dict with tensor values
# The function expects dict values to be tensors that can be stacked and reduced
input_dict = {
    "loss": torch.tensor(0.5),
    "accuracy": torch.tensor(0.95),
    "metric_c": torch.tensor(2.3),
}

ARGS = {
    "input_dict": input_dict,
    "average": True,
    "to_item": True,
}
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 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
upcast Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/curve_utils.py
code served (permissive licence) · get_code("944efb6b2d124909")
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 1, dtype family float, shape (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)

# The function 'upcast' takes a single parameter 't' which is a tensor
# It checks if the tensor is floating point or integer type
# and upcasts it to float32/float64 or int32/int64 respectively

# Test with a floating point tensor (float16 - needs upcasting to float32)
t_float16 = torch.tensor([1.0, 2.0, 3.0, 4.0], dtype=torch.float16)

ARGS = {
    "t": t_float16
}
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 1, dtype family float, shape (4,) — shape (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
update_nested Ran
ran on a synthesized input
voldemortX/pytorch-auto-drive/utils/args.py
code served (permissive licence) · get_code("6d390a74f15c1b8d")
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 torch

# Construct valid arguments for update_nested function
# The function takes a dict, a list of keys (path), and a value
# It recursively updates nested dictionaries

# Create a nested dictionary
d = {
    "level1": {
        "level2": {
            "level3": "old_value"
        },
        "other_key": "other_value"
    },
    "simple_key": "simple_value"
}

# Create a list of keys representing a path into the nested dict
keys = ["level1", "level2", "level3"]

# Create a value to set
value = "new_value"

ARGS = {
    "d": d,
    "keys": keys,
    "value": value
}
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 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

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 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
counter_conv 1 other paper
1 later
2211.06088 (2022-11)
RepGhost: A Hardware-Efficient Ghost Module via Re-parameterization
reduce_dict 1 other paper
1 later
2405.06264 (2024-05)
Selective Focus: Investigating Semantics Sensitivity in Post-training Quantization for Lane Detection

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

This paper presents a novel parametric curve-based method for lane detection in RGB images. Unlike stateof-the-art segmentation-based and point detection-based methods that typically require heuristics to either decode predictions or formulate a large sum of anchors, the curvebased methods can learn holistic lane representations naturally. To handle the optimization difficulties of existing polynomial curve methods, we propose to exploit the parametric Bézier curve due to its ease of computation, stability, and high freedom degrees of transformations. In addition, we propose the deformable convolution-based feature flip fusion, for exploiting the symmetry properties of lanes in driving scenes. The proposed method achieves a new state-ofthe-art performance on the popular LLAMAS benchmark. It also achieves favorable accuracy on the TuSimple and CULane datasets, while retaining both low latency (>150 FPS) and small model size (<10M). Our method can serve as a new baseline, to shed the light on the parametric curves modeling for lane detection. Codes of our model and PytorchAutoDrive: a unified framework for self-driving perception, are available at: https://github.com/ voldemortX/pytorch-auto-drive .

For agents

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

get_harvested_code_for_paper("2203.02431")
get_code_for_paper("2203.02431")
have("2203.02431")

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

curl https://syntology.ai/api/ran/2203.02431.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.02431.svg)](https://syntology.ai/paper/2203.02431)

Connect an agent — have() is free.