SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2107.09584 · 2021

Active 3D Shape Reconstruction from Vision and Touch

arXiv · PDF · Open in the Atlas

Code that ran

5 of 7 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 4 have no contract verdict; 1 failure of our instrument; 2 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/2107.09584.json · badge: /api/ran/2107.09584.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
facebookresearch/Active-3D-Vision-and-Touch canonical (classifier label)
classifier basis: census
3 ran · 0 no recorded run
maurock/object_reconstruction community (classifier label)
classifier basis not recorded on the link
2 ran · 2 no recorded run
FunctionStatusWhere it lives
CNN_layer Ran
ran on a synthesized input
facebookresearch/Active-3D-Vision-and-Touch/pterotactyl/reconstruction/vision/model.py
code served (permissive licence) · get_code("228ad0dc3534890a")
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 = {
    'f_in': 3,
    'f_out': 16,
    'k': 3,
    'stride': 1,
    'simple': False,
    'padding': 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
GridSamplingLayer Ran
ran; our contract draft was wrong, not the code
facebookresearch/Active-3D-Vision-and-Touch/pterotactyl/reconstruction/autoencoder/model.py
code served (permissive licence) · get_code("d734b3d4a40b8639")
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 (2, 24, 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

# GridSamplingLayer parameters:
# - batch_size: int, number of batches
# - meshgrid: list of tuples, each tuple is (min, max, num_points)

batch_size = 2

# Create a simple meshgrid specification
# Each element is (min_val, max_val, num_points)
meshgrid = [
    (0.0, 1.0, 3),    # First dimension: 3 points from 0 to 1
    (-1.0, 1.0, 4),   # Second dimension: 4 points from -1 to 1
    (0.0, 2.0, 2),    # Third dimension: 2 points from 0 to 2
]

ARGS = {
    "batch_size": batch_size,
    "meshgrid": meshgrid,
}
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 (2, 24, 3) — shape (2, 24, 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
prepare_mesh Ran
ran on a synthesized input
facebookresearch/Active-3D-Vision-and-Touch/pterotactyl/reconstruction/vision/model.py
code served (permissive licence) · get_code("5bd483228b9a6acf")
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. 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 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 a small batch dict with 'img' and 'touch_charts'
batch = {
    'img': torch.randn(2, 3, 32, 32),  # batch_size=2
    'touch_charts': torch.randn(2, 5, 4)  # batch_size=2, 5 touch points, 4 channels (3 coords + 1 mask)
}

# Create a small vision_mesh tensor
# The function does vision_mesh.unsqueeze(0).repeat(s1, 1, 1)
# So vision_mesh should be 2D: (num_points, 3)
vision_mesh = torch.randn(7, 3)

# Create args object with use_touch attribute
class Args:
    def __init__(self, use_touch):
        self.use_touch = use_touch

args_with_touch = Args(use_touch=True)
args_without_touch = Args(use_touch=False)

# Test with use_touch=True
ARGS = {
    'batch': batch,
    'vision_mesh': vision_mesh,
    'args': args_with_touch
}
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
robot_touch_vertical Ran
ran on a synthesized input
maurock/object_reconstruction/object_reconstruction/utils/samples_utils.py
code served (permissive licence) · get_code("1c64dc1d2c29a88e")
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 MockRobot. 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 time

class MockArm:
    def worldframe_to_workframe(self, position, orientation):
        """Convert world frame to work frame coordinates"""
        # Simple mock: just return the position as work frame with some offset
        work_position = np.array(position) + np.array([0.1, 0.1, 0.05])
        return work_position, orientation

class MockRobot:
    def __init__(self):
        self.arm = MockArm()
        self.stop_at_touch = False
        self.last_position = None
        self.last_orientation = None
    
    def move_linear(self, position, orientation):
        """Mock linear movement"""
        self.last_position = position
        self.last_orientation = orientation
        # Simulate movement delay
        time.sleep(0.01)

# Create robot instance
robot = MockRobot()

# Create sample coordinates (x, y) in world frame
sample = [0.5, 0.3]

# Create ARGS dictionary
ARGS = {
    "robot": robot,
    "sample": sample,
    "max_height_wrld": 0.2
}
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 MockRobot — a MockRobot output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockRobot 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
sample_hemisphere Ran
ran on a synthesized input
maurock/object_reconstruction/object_reconstruction/utils/samples_utils.py
code served (permissive licence) · get_code("d25800c157d0dd64")
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. 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

# Set random seed for reproducibility
np.random.seed(0)

# The function sample_hemisphere takes one parameter: r (radius)
# It uses r as a scalar to compute coordinates on a hemisphere
# Looking at the function body:
# - r is multiplied with sin(theta), cos(phi), etc. to get x, y, z coordinates
# - r should be a numeric scalar (float or int)

ARGS = {
    "r": 5.0
}
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 tuple — a tuple 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
generate_touch_vision_data Not yet run
unverified (harvested, no recorded run)
maurock/object_reconstruction/object_reconstruction/data_making/extract_deformation.py
code served (permissive licence) · get_code("775d896091ceaeae")
How it ran

no invocation recorded.

sphere_orn_wrld Not yet run
unverified (harvested, no recorded run)
maurock/object_reconstruction/object_reconstruction/utils/samples_utils.py
code served (permissive licence) · get_code("77ed3637ce720715")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 6 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
GridSamplingLayer 1 other paper
1 dated earlier
1612.00593 (2016-12)
PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation

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

Humans build 3D understandings of the world through active object exploration, using jointly their senses of vision and touch. However, in 3D shape reconstruction, most recent progress has relied on static datasets of limited sensory data such as RGB images, depth maps or haptic readings, leaving the active exploration of the shape largely unexplored. Inactive touch sensing for 3D reconstruction, the goal is to actively select the tactile readings that maximize the improvement in shape reconstruction accuracy. However, the development of deep learning-based active touch models is largely limited by the lack of frameworks for shape exploration. In this paper, we focus on this problem and introduce a system composed of: 1) a haptic simulator leveraging high spatial resolution vision-based tactile sensors for active touching of 3D objects; 2)a mesh-based 3D shape reconstruction model that relies on tactile or visuotactile signals; and 3) a set of data-driven solutions with either tactile or visuotactile priors to guide the shape exploration. Our framework enables the development of the first fully data-driven solutions to active touch on top of learned models for object understanding. Our experiments show the benefits of such solutions in the task of 3D shape understanding where our models consistently outperform natural baselines. We provide our framework as a tool to foster future research in this direction.

For agents

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

get_harvested_code_for_paper("2107.09584")
get_code_for_paper("2107.09584")
have("2107.09584")

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

curl https://syntology.ai/api/ran/2107.09584.json

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

[![Syntology run record](https://syntology.ai/api/ran/2107.09584.svg)](https://syntology.ai/paper/2107.09584)

Connect an agent — have() is free.