SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2409.00768 · ECCV · 2024

Rethinking Image Super-Resolution from Training Data Perspectives

Yuki Asano, Christian Rupprecht, Iro Laina, Nakamasa Inoue, Rio Yokota, Hirokatsu Kataoka, Yoshimitsu Aoki, Ryu Tadokoro, Ryosuke Yamada, Go Ohtani

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures 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/2409.00768.json · badge: /api/ran/2409.00768.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
gohtanii/DiverSeg-dataset marked official in Papers with Code
Papers with Code archive
4 ran · 0 no recorded run
FunctionStatusWhere it lives
calc_DCT Ran
ran; our contract draft was wrong, not the code
gohtanii/DiverSeg-dataset/blockiness/calc_blockiness.py
code served (permissive licence) · get_code("4f54d60e12da1bbb")
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 (16, 16) 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
from scipy import fftpack

class DCTWrapper:
    """Minimal DCT wrapper that implements the dct2 method"""
    def dct2(self, block):
        """2D Discrete Cosine Transform"""
        return fftpack.dct(fftpack.dct(block.T, norm='ortho').T, norm='ortho')

# Create test image (16x16 to work with 8x8 blocks)
# Using 2 blocks horizontally and 2 blocks vertically
img = np.random.randn(16, 16).astype(np.float32)

# DCT object with dct2 method
dct = DCTWrapper()

# Number of blocks
h_block_num = 2
w_block_num = 2

ARGS = {
    'img': img,
    'dct': dct,
    'h_block_num': h_block_num,
    'w_block_num': w_block_num
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 (16, 16) — shape (16, 16) 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
calc_V Ran
ran; our contract draft was wrong, not the code
gohtanii/DiverSeg-dataset/blockiness/calc_blockiness.py
code served (permissive licence) · get_code("ae271c1df44dd73e")
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 (8, 8) 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

# Analyze the function to determine required arguments
# calc_V takes:
# - dct_img: a 2D numpy array that is indexed with [h_idx, w_idx]
# - h_block_num: number of blocks in height dimension
# - w_block_num: number of blocks in width dimension

# The function accesses indices like:
# w_idx = BLOCK_SIZE + w_block * BLOCK_SIZE + i  (ranges: BLOCK_SIZE to BLOCK_SIZE + (w_block_num-3)*BLOCK_SIZE + 7)
# h_idx = BLOCK_SIZE + h_block * BLOCK_SIZE + j  (ranges: BLOCK_SIZE to BLOCK_SIZE + (h_block_num-3)*BLOCK_SIZE + 7)
# with w_block in range(1, w_block_num - 2) and h_block in range(1, h_block_num - 2)
# Also accesses w_idx ± BLOCK_SIZE and h_idx ± BLOCK_SIZE

# So we need:
# max w_idx = BLOCK_SIZE + (w_block_num-3)*BLOCK_SIZE + 7
# max h_idx = BLOCK_SIZE + (h_block_num-3)*BLOCK_SIZE + 7
# We also need w_idx - BLOCK_SIZE >= 0 and w_idx + BLOCK_SIZE to be valid
# And h_idx - BLOCK_SIZE >= 0 and h_idx + BLOCK_SIZE to be valid

BLOCK_SIZE = 8

# Let's use h_block_num = 5 and w_block_num = 5
# Then w_block ranges from 1 to 2 (1, w_block_num-2 = 3, so range(1,3) = [1,2])
# max w_idx = BLOCK_SIZE + 2*BLOCK_SIZE + 7 = 8 + 16 + 7 = 31
# Similarly for h_idx = 31
# We need w_idx + BLOCK_SIZE = 31 + 8 = 39, so width must be at least 40
# Similarly height must be at least 40

h_block_num = 5
w_block_num = 5

# Create dct_img with sufficient size
# Height needs to be at least 40, width at least 40
dct_img = np.random.randn(40, 40)

ARGS = {
    'dct_img': dct_img,
    'h_block_num': h_block_num,
    'w_block_num': w_block_num
}
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 (8, 8) — shape (8, 8) 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
calc_margin Ran
ran, honoured the contract we drafted
gohtanii/DiverSeg-dataset/blockiness/calc_blockiness.py
code served (permissive licence) · get_code("a2d2ec907298953a")
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) and returned tuple, ndim 1, dtype family int; 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

BLOCK_SIZE = 8

def calc_margin(height, width):
    h_margin = height % BLOCK_SIZE
    w_margin = width % BLOCK_SIZE
    cal_height = height - (h_margin if h_margin >= 4 else h_margin + BLOCK_SIZE)
    cal_width = width - (w_margin if w_margin >= 4 else w_margin + BLOCK_SIZE)
    h_margin = h_margin + BLOCK_SIZE if h_margin < 4 else h_margin
    w_margin = w_margin + BLOCK_SIZE if w_margin < 4 else w_margin
    return (cal_height, cal_width, h_margin, w_margin)

ARGS = {
    "height": 25,
    "width": 33
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 int — 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
process_image Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
gohtanii/DiverSeg-dataset/blockiness/calc_blockiness.py
code served (permissive licence) · get_code("6f93b531ec14ffeb")
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) and returned float64, ndim 0, 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 numpy as np
import cv2

class DCTMock:
    """Mock DCT object with dct2 method"""
    def dct2(self, block):
        """Simple mock DCT that returns a transformed block"""
        from scipy.fftpack import dct
        return dct(dct(block, axis=0, norm='ortho'), axis=1, norm='ortho')

# Create a small test image (BGR format, 3 channels)
# Using 32x32 to ensure we have enough blocks for the algorithm
# BLOCK_SIZE = 8, so 32x32 gives us 4x4 blocks
test_img = np.random.randint(0, 256, (32, 32, 3), dtype=np.uint8)

# Create DCT mock object
dct_mock = DCTMock()

ARGS = {
    "img": test_img,
    "dct": dct_mock
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
cv2, numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 4 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.

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

In this work, we investigate the understudied effect of the training data used for image super-resolution (SR). Most commonly, novel SR methods are developed and benchmarked on common training datasets such as DIV2K and DF2K. However, we investigate and rethink the training data from the perspectives of diversity and quality, thereby addressing the question of "How important is SR training for SR models?". To this end, we propose an automated image evaluation pipeline. With this, we stratify existing high-resolution image datasets and larger-scale image datasets such as ImageNet and PASS to compare their performances. We find that datasets with (i) low compression artifacts, (ii) high within-image diversity as judged by the number of different objects, and (iii) a large number of images from ImageNet or PASS all positively affect SR performance. We hope that the proposed simple-yet-effective dataset curation pipeline will inform the construction of SR datasets in the future and yield overall better models. Code

For agents

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

get_harvested_code_for_paper("2409.00768")
get_code_for_paper("2409.00768")
have("2409.00768")

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

curl https://syntology.ai/api/ran/2409.00768.json

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

[![Syntology run record](https://syntology.ai/api/ran/2409.00768.svg)](https://syntology.ai/paper/2409.00768)

Connect an agent — have() is free.