Binocular Perception Instance Authentication Learning for Few-Shot Visual Recognition.
3 blocks of code here have a permanent identifier (oscr:b:…), listed with the matches below: citable for good, with a page of its own. Select a match and press Cite to copy it. What a block identifier is.
The 4 matches · 2 of them tie a paragraph to a whole file without given lines: weak matches, whose lines are not tinted
- [1] § 4. Experiments › 4.2. Implementation Details ↔ config.py, the whole file · a weak match · score 0.93 · Aircraft Fewshot, weight decay, decay step, TieredImagenet, miniImageNet, FS · block
oscr:b:410XAZ3J3THT3-Z· Report a mismatch - [2] § 4. Experiments › 4.4. Ablation Study ↔ config.py, the whole file · a weak match · score 0.65 · Expand Ratio, Confidence Ratio, TieredImagenet, algorithm, BPIAL, class · block
oscr:b:410XAZ3J3THT3-Z· Report a mismatch - [3] § 4. Experiments › 4.2. Implementation Details ↔ main.py, lines 45 to 138 · score 0.63 · weight decay, meta training, epochs, optimizer, benchmarks, accuracy · block
oscr:b:TP2YSJ787GQ1F-Y· Report a mismatch - [4] § 3. Methodology › 3.3. Binocular Sensing Extractor Module (BSEM) ↔ models/vision_res12_le.py, lines 124 to 183 · score 0.52 · vision layer, vision res12, batch, Module · block
oscr:b:NTT30E5RGDV63-N· Report a mismatch
Paper
Loaded from Europe PMC by your browser; OSCR does not store it: doi.org · Europe PMC
The paper is loaded when this pane is shown.
The authors' code
Python · 120 lines · 7.2 KB · CC-BY-NC-4.0 · 2 matches
- # +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
- # Author: CHAOFEI QI
- # Email: [email hidden]
- # Address: Harbin Institute of Technology
- #
- # Copyright (c) 2024
- # This source code is licensed under the MIT-style license found in the
- # LICENSE file in the root directory of this source tree
- # +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
- import argparse
- def config():
- parser = argparse.ArgumentParser(description='Train image model with cross entropy loss')
- # ************************************************************
- # Datasets (general)
- # ************************************************************
- parser.add_argument('-d', '--dataset', type=str, default='miniImageNet')
- parser.add_argument('--load', default=False)
- parser.add_argument('-j', '--workers', default=8, type=int, help="number of data loading workers (default: 4)") # 11-2
- parser.add_argument('--height', type=int, default=84, help="height of an image (default: 84)")
- parser.add_argument('--width', type=int, default=84, help="width of an image (default: 84)")
- # ************************************************************
- # Miscs
- # ************************************************************
- parser.add_argument('--scale_cls', type=int, default=7)
- parser.add_argument('--save-dir', type=str)
- parser.add_argument('--resume', type=str, default=None, metavar='PATH')
- parser.add_argument('-g', '--gpu-devices', default='0', type=str)
- parser.add_argument('--checkpoint', default=False)
- # ************************************************************
- # BPIAL parameters
- # ************************************************************
- parser.add_argument('--classifier', type=str, default='lr')
- parser.add_argument('--disparity', type=str, default='js')
- parser.add_argument("--T", type=float, default=4.0)
- parser.add_argument('--lembed', type=str, default='se')
- parser.add_argument('--rembed', type=str, default='se')
- parser.add_argument('--emb_dim', type=int, default=5)
- parser.add_argument('--logit_penalty', type=float, default=0.5)
- parser.add_argument('--confidence_ratio', type=float, default=0.6)
- parser.add_argument("--method", type=str, choices=['1','2','3'], default='2') # Input method
- parser.add_argument("--FFT_sign", default=False) # Input method
- parser.add_argument("--Elastic", type=str, default='1', choices=['0','1'], help='elastic_constraint') # [0(Flase),1(True)]
- parser.add_argument("--weights", type=str, default="1-1-1")
- parser.add_argument("--fit_mode", type=str, default="db", choices=['db'])
- parser.add_argument('--expand_ratio', type=float, default=0.5)
- parser.add_argument("--stereopsis", type=str, default="intersection", choices=['intersection','union','le','re'])
- parser.add_argument("--expand_style", type=str, default="linear", choices=['linear'])
- # ************************************************************
- # FewShot settting
- # ************************************************************
- parser.add_argument('--mode', type=str, default='train')
- parser.add_argument('--phase', default='test', type=str, help='use test or val dataset to early stop')
- parser.add_argument('--n_train_ways', type=int, default=15, metavar='N', help='Number of classes for doing each classification run')
- parser.add_argument('--train_nTestNovel', type=int, default=7 * 15, help='number of test examples for all the novel category when training')
- parser.add_argument('--nKnovel', type=int, default=5, help='number of novel categories')
- parser.add_argument('--meta_nTestNovel', type=int, default=7 * 5, help='number of test examples for all the novel category when training')
- parser.add_argument('--nExemplars', type=int, default=5, help='number of training examples per novel category.')
- parser.add_argument('--unlabel', type=int, default=0)
- parser.add_argument('--nTestNovel', type=int, default=15 * 5, help='number of test examples for all the novel category')
- parser.add_argument('--start-epoch', default=0, type=int, help="manual epoch number (useful on restarts)")
- parser.add_argument('--max-epoch', default=90, type=int, help="maximum epochs to run")
- parser.add_argument('--stepsize', default=[60], nargs='+', type=int, help="stepsize to decay learning rate")
- parser.add_argument('--train_epoch_size', type=int, default=1200, help='number of batches per epoch when training')
- parser.add_argument('--epoch_size', type=int, default=2000, help='number of batches per epoch')
- parser.add_argument('--test-batch', default=1, type=int, help="test batch size")
- parser.add_argument('--train-batch', default=2, type=int, help="train batch size") # 11-11
- # ************************************************************
- # Optimization options
- # ************************************************************
- parser.add_argument('--optim', type=str, default='SGD', help="optimization algorithm (see optimizers.py)") # sgd,SGD
- parser.add_argument('--lr', '--learning-rate', default=0.1, type=float, help="initial learning rate") # 初始学习率
- parser.add_argument("--decay_step", type=int, default=40)
- parser.add_argument('--LUT_lr', default=[(60, 0.1), (80, 0.006), (90, 0.0012)], help="multistep to decay learning rate") # optim='sgd'时使用
- parser.add_argument('--weight-decay', default=5e-04, type=float, help="weight decay (default: 5e-04)")
- # ************************************************************
- # DistributedDataParallel
- # ************************************************************
- parser.add_argument('--amp_opt_level', type=str, default='O0', choices=['O0', 'O1', 'O2'], help='mixed precision opt level, if O0, no amp is used')
- parser.add_argument("--seed", default='1', type=str)
- # ***********************************************************
- args = parser.parse_args()
- args.w = [float(i) for i in args.weights.split('-')]
- args.train_epoch_size= int(600*args.train_batch)
- if args.dataset=='CIFARFS':
- args.num_classes = 64
- args.dataset_dir = '/home/ssdData/qcfData/fs_benchmarks/CIFARFS/'
- elif args.dataset=='CUB_Croped':
- args.num_classes = 100
- args.dataset_dir = '/home/ssdData/qcfData/fs_benchmarks/CUB_fewshot_cropped'
- elif args.dataset=='Aircraft':
- args.num_classes = 50
- args.dataset_dir = '/home/ssdData/qcfData/fs_benchmarks/Aircraft_fewshot/'
- elif args.dataset=='TieredImagenet':
- args.num_classes = 351
- args.train_epoch_size = 13980
- args.max_epoch = 120
- args.LUT_lr = [(30, 0.1), (60, 0.01), (90, 0.001),(120, 0.0001)]
- args.dataset_dir = '/home/ssdData/qcfData/fs_benchmarks/TieredImagenet_224'
- elif args.dataset=='MiniImagenet':
- args.num_classes = 64
- args.dataset_dir = '/home/ssdData/qcfData/fs_benchmarks/mini_imagenet_full_size/mini_imagenet_full_size'
- else:
- print("No dataset supported.")
- return args
config.py at commit 6257dff, under CC-BY-NC-4.0 · at the source
Overview
- Faculty of Computing, Harbin Institute of Technology, Harbin 150001, China
- School of Mechanical Engineering and Automation, Harbin Institute of Technology Shenzhen, Shenzhen 518055, China
- Research Institute of Intelligent Control and Systems, Harbin Institute of Technology, Harbin 150001, China
Abstract
Humans possess unique advantages in dealing with few-shot visual recognition scenarios, inspiring the development of meta-learning methods aimed at emulating these abilities. Current mainstream meta-learning primarily utilizes the monocular vision or the dual asymmetric complementary architectures, collectively referred to as Machine-visual Meta-Learning (MvML). Nevertheless, research has not yet developed an architecture for simulating the human binocular visual system, named Humanoid-visual Meta-Learning (HvML). This paper innovatively proposes an excellent paradigm BPIAL belonging to the HvML: Binocular Perception Instance Authentication Learning, which can alleviate the monocular shallowness and dual-branch processing instability of MvML. Structurally, our BPIAL comprises two interconnected binocular perception and information processing modules: BSEM and IAPM. The former module can simulate binocular visual field extraction, feature extraction and compression, and channel dimension reduction, while IAPM can simulate the logic, reasoning, and judgment processes of the two human visual branches and the brain. We have demonstrated the feasibility and correctness of BPIAL on five benchmarks. Sufficient and comparative experiments with the state-of-the-art methods have proved its superiority and effectiveness.
Reproduced under the paper's license (CC BY), from the paper cited above.
Repository
Its files are read in the Code ↔ Paper reader above, with 4 matches between paragraphs and lines of code.
ChaofeiQI/BPIAL
6257dffa14c0f77d4eefec363ff86d7921f6d14a, 30 July 2025Availability: 1 check, the latest on 27 September 2026: the link answers
- 27 September 2026: the link answers
36 files
- config.py, Python, 120 lines, 2 matches
- data/
__init__.py , Python, 1 line - data/
data_manager.py , Python, 193 lines - data/
loaders/ , Python, 22 lines__init__.py - data/
loaders/ , Python, 159 linestest_loader.py - data/
loaders/ , Python, 137 linestrain_loader.py - data/
loaders/ , Python, 130 linestrain_loader_mini.py - data/
sets/ , Python, 24 lines__init__.py - data/
sets/ , Python, 64 linesaircraft.py - data/
sets/ , Python, 64 linescifarfs.py - data/
sets/ , Python, 64 linescub_croped.py - data/
sets/ , Python, 64 linesminiImageNet.py - data/
sets/ , Python, 64 linestieredImageNet.py - exp_test.sh, Shell, 22 lines
- exp_train.sh, Shell, 19 lines
- main.py, Python, 315 lines, 1 match
- models/
BSEM.py , Python, 311 lines - models/
IAPM.py , Python, 236 lines - models/
__init__.py , Python, 1 line - models/
backbones/ , Python, 212 linesbigres12.py - models/
backbones/ , Python, 32 linesconvnet.py - models/
backbones/ , Python, 61 linesdropblock.py - models/
backbones/ , Python, 211 linesnewres12.py - models/
vision_res12_le.py , Python, 189 lines, 1 match - models/
vision_res12_re.py , Python, 188 lines - utils/
__init__.py , Python, 1 line - utils/
avgmeter.py , Python, 23 lines - utils/
ci.py , Python, 9 lines - utils/
iotools.py , Python, 46 lines - utils/
logger.py , Python, 45 lines - utils/
losses.py , Python, 43 lines - utils/
optimizers.py , Python, 49 lines - utils/
torchtools.py , Python, 93 lines - utils/
transforms.py , Python, 90 lines - LICENSE.txt, License, 408 lines
- README.md, Text, 136 lines
The paper's code and data availability statement is in the Data section.
Tracing map
Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.
What the map holds:
- 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
- 34 scripts, each with its path and the digest of its content;
- 4 matches between paragraphs of the paper and lines of the code (method lexical-v1);
- neither the text of the paper nor the code itself.
Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.
Data
No dataset and no data link were found in the paper.
Data Availability Statement
The data is available at https://
Reproduced under the paper's license (CC BY), from the paper cited above.
Versions
The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.
Version 1, 27 September 2026: the first record
Recorded: type, language, journal, volume, issue, pages, dates, 3 authors, 5 keywords, 2 funders, 49 references.
Cite
This paper
Qi, C., Li, P., & Lin, W. (2026). Binocular Perception Instance Authentication Learning for Few-Shot Visual Recognition. Biomimetics (Basel, Switzerland), 11(7), 505. https://
BibTeX
@article{qi2026binocular
author = {Qi, Chaofei and Li, Peng and Lin, Weiyang},
title = {{Binocular Perception Instance Authentication Learning for Few-Shot Visual Recognition}},
journal = {Biomimetics (Basel, Switzerland)},
year = {2026},
month = jul,
volume = {11},
number = {7},
pages = {505},
publisher = {Multidisciplinary Digital Publishing Institute (MDPI)},
issn = {2313-7673},
doi = {10.3390/
url = {https://
pmid = {42505538},
pmcid = {PMC13406620}
}
RIS
TY - JOUR
AU - Qi, Chaofei
AU - Li, Peng
AU - Lin, Weiyang
TI - Binocular Perception Instance Authentication Learning for Few-Shot Visual Recognition
T2 - Biomimetics (Basel, Switzerland)
J2 - Biomimetics (Basel)
PY - 2026
DA - 2026/
VL - 11
IS - 7
SP - 505
SN - 2313-7673
PB - Multidisciplinary Digital Publishing Institute (MDPI)
DO - 10.3390/
UR - https://
LA - en
ER -
CSL-JSON
{
"id": "10.3390/
"type": "article-journal",
"title": "Binocular Perception Instance Authentication Learning for Few-Shot Visual Recognition",
"container-title": "Biomimetics (Basel, Switzerland)",
"author": [
{
"family": "Qi",
"given": "Chaofei"
},
{
"family": "Li",
"given": "Peng"
},
{
"family": "Lin",
"given": "Weiyang"
}
],
"container-title-short":
"volume": "11",
"issue": "7",
"page": "505",
"DOI": "10.3390/
"PMID": "42505538",
"PMCID": "PMC13406620",
"ISSN": "2313-7673",
"publisher": "Multidisciplinary Digital Publishing Institute (MDPI)",
"URL": "https://
"language": "en",
"issued": {
"date-parts": [
[
2026,
7,
18
]
]
}
}
The tracing map gets a citation of its own once an author has validated it and it has a DOI.
Similar papers
The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.
- [1] doi:10.1038/s41598-025-18329-8 [code]
- Multi-stage fusion of local and global features for few-shot image classification.Journal: Scientific reportsIn common: Pillow, PyTorch, scikit-learn, 1 other tool, 5 references
- [2] doi:10.1371/journal.pcbi.1014656 [code]
- Contrastive learning to fine-tune feature extraction models for the visual cortex.Journal: PLoS computational biologyIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [3] doi:10.1038/s41467-026-69345-9 [code]
- A 7T fMRI dataset of synthetic images for out-of-distribution modeling of vision.Journal: Nature communicationsIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [4] doi:10.1038/s41597-025-05174-7 [code]
- A large-scale MEG and EEG dataset for object recognition in naturalistic scenesJournal: n/aIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [5] doi:10.1371/journal.pcbi.1013391 [code]
- A feedforward mechanism for human-like contour integration.Journal: PLoS computational biologyIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [6] doi:10.1038/s41562-025-02252-z [code]
- In silico discovery of representational relationships across visual cortex.Journal: Nature human behaviourIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [7] doi:10.7554/elife.96217 [code]
- Convolutional networks can model the functional modulation of the MEG responses associated with feed-forward processes during visual word recognition.Journal: eLifeIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [8] doi:10.3389/fninf.2025.1515873 [code]
- Net2Brain: a toolbox to compare artificial vision models with human brain responses.Journal: Frontiers in neuroinformaticsIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [9] doi:10.1038/s41598-024-80938-6 [code]
- Enhancement and evaluation for deep learning-based classification of volumetric neuroimaging with 3D-to-2D knowledge distillation.Journal: Scientific reportsIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
- [10] doi:10.1038/s41598-024-66228-1 [code]
- Image classification and reconstruction from low-density EEG.Journal: Scientific reportsIn common: Pillow, PyTorch, scikit-learn, 2 other tools, 1 reference
Contribute
The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.
Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.
Claim this paper
Correct its record
Say what each link of this record is, remove the ones that are not the paper's, add the ones that are missing. The correction becomes a new version of the record, in its Versions section.
Validate its tracing map
You validate the map as this page shows it: 1 repository of the authors' code, each at its verified commit and with its license, 34 scripts, and 4 matches between paragraphs and code (see the Code and Map sections). It then receives a DOI on Zenodo, with you (your ORCID iD) and OSCR as its creators; the code itself is not deposited.
The map's fingerprint: sha256:2907528fd9e100d6…
Add the badge to the code's README
If you are an author or a maintainer of this paper's code, you can add this badge. It says what OSCR knows of the paper (its code, its validated tracing map) and links the code to this page. Paste one of these snippets under the README's title. Keep the two marks around it, so that the badge can be replaced in place and is never added twice. OSCR changes nothing in your repository.
Markdown
<!-- oscr-badge:start -->
[](https://openscicode.org/badge/doi/10.3390/biomimetics11070505)
<!-- oscr-badge:end -->reStructuredText
.. oscr-badge:start
.. image:: https://openscicode.org/badge/doi/10.3390/biomimetics11070505.svg
:target: https://openscicode.org/badge/doi/10.3390/biomimetics11070505
:alt: OSCR
.. oscr-badge:endHTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.3390/biomimetics11070505"><img src="https://openscicode.org/badge/doi/10.3390/biomimetics11070505.svg" alt="OSCR"></a>
<!-- oscr-badge:end -->On GitHub, open the README, choose Edit (the pencil), paste the snippet under the title, then "Commit changes".
If you choose to, you can have the badge kept up to date for you, with oscr badge --write from the command line or with the GitHub Action in your repository's workflows. Every other form and option (a card, AsciiDoc, LaTeX, shields.io) is on the badge page.
Request its removal
To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).
Discussion, reproductions, activity
Discussion: code errors and mismatches between the code and the paper about this paper and its code, from readers and its authors, beside the paper and its tracing map.
Report a mismatch between the code and the paper · Report a code error (signed in)
Reproductions: reports from readers who ran the authors' code and did not get the paper's result, what they ran, with which environment, commit and data.
Report a reproduction failure (signed in)
Versions of its code: the releases of its code tied to a version of this paper (the preprint, the accepted manuscript, the version of record…), each with the tracing map versioned with it.
Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.
