Pith. sign in

REVIEW 4 major objections 6 minor 30 references

SEMU-Net: A Segmentation-based Corrector for Fabrication Process Variations of Nanophotonics with Microscopic Images

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read SEMU-Net claims a segmentation U-Net plus a tandem attention-U-Net corrector can make fabricated nanophotonic structures match intended GDS designs to an average IoU of 98.67% on a custom benchmark of stars, gratings, circles, and crosses.

desk verdict The segmentation experiment is plausibly real, but the 98.67% corrector IoU is a closed-loop score against the authors' own predictor, and without a fabricated corrected layout the paper's central claim does not follow. read the letter →

arxiv 2411.16973 v1 pith:VKWQIXZY submitted 2024-11-25 cs.CV eess.IV

classification cs.CVeess.IV
keywords siliconphotonicsSEMimagesegmentationU-Netfabricationprocessvariationsinversedesignattentiontandemarchitectureintersectionoverunion
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that nanofabrication deviations in silicon photonics—over-etching, under-etching, and corner rounding—are learnable patterns, and that a three-stage deep learning pipeline can compensate for them before fabrication. It introduces SEMU-Net: a U-Net that segments scanning electron microscope images of fabricated devices into silicon and silica masks, a predictor that learns the mapping from design file (GDS) to fabricated shape, and a corrector that learns the reverse mapping to adjust design files. The key result is that the corrector, an attention U-Net trained in tandem with a frozen predictor, reaches an average IoU of 98.67% on a custom benchmark, meaning corrected layouts are predicted to fabricate almost exactly as intended. If correct, this would let designers send pre-compensated layouts to the foundry and get devices closer to specification without repeated redesign cycles.

What carries the argument

The central mechanism is the tandem architecture: a corrector (attention U-Net) is stacked with a predictor (U-Net with frozen weights), and the corrector is trained to minimize the discrepancy between the predictor's output on the corrected GDS and the original design, so the corrector learns an identity mapping from GDS to GDS through the fabrication-distortion bottleneck. The attention gate refines decoder features by weighting them with a gating signal from the encoder, focusing correction where fabrication shifts the geometry. The segmentation U-Net supplies the ground-truth masks that define what 'fabricated' means in the training loop.

What would settle it

Fabricate the corrected GDS layouts, image them with SEM, and compare the measured structures against the intended designs using an independent segmentation method or direct optical characterization. A large drop in IoU relative to the 98.67% predicted would show the corrector is compensating for segmentation artifacts rather than real process variations. Even simpler: have human experts manually annotate a held-out set of SEM images and re-run the corrector evaluation against those annotations.

Watch

Extended reading notes

Core claim

The central claim is that fabrication-induced structural variations in nanophotonic devices can be automatically segmented from SEM images and then corrected by learning an inverse mapping from fabricated appearance back to the design file. The authors show that a U-Net segmenter recognizes silicon versus silica at a 99.30% average IoU, and that a tandem attention U-Net corrector, evaluated by its predicted post-fabrication shape against the original design on a benchmark of stars, gratings, circles, and crosses, achieves a 98.67% average IoU. The corrector effectively produces GDS layouts with deliberate compensation such that, after the predicted fabrication distortions, the structure matches the intended design.

Load-bearing premise

The load-bearing premise is that the segmented SEM images used as ground truth faithfully represent the real fabricated geometry; the whole pipeline is trained and scored against these segmentations, so any bias in the segmenter would be learned as if it were a fabrication effect, and the reported IoU could overstate true alignment with the physical device.

Editorial extensions

If this is right

  • A foundry could use SEMU-Net as a pre-fabrication step: run the predictor on a proposed layout, run the corrector, and send the corrected GDS to manufacturing, reducing reliance on simulation-based biasing.
  • The same pipeline could generalize to other fabrication processes—photolithography, etching, or deposition—where a forward model or image of the fabricated result is available for training.
  • Because the corrector is evaluated by its own predicted post-fabrication shape, the method enables in-loop validation: designers can check whether the corrected layout will survive fabrication before spending chip area.
  • The high IoU on shapes like stars and gratings suggests the method captures localized effects such as corner rounding, not just uniform bias, which simulation-based approaches may miss.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If fabrication variations are consistent across runs on the same platform, the trained corrector could be reused for many designs with minimal retraining, converting a per-layout correction step into a one-time process characterization.
  • The tandem loop suggests a general recipe for any fabrication flow with a differentiable or learnable forward model: train a corrector by backpropagating through the frozen forward model, which could apply beyond SOI waveguides to other nanofabrication steps.
  • Since the paper benchmarks geometric overlap on a few shape classes, an optical test on corrected devices—measuring transmission spectra or resonance wavelengths—would be a stronger validation, because devices with equal IoU can still have different electromagnetic responses.
  • The reported 99.30% segmentation IoU might not transfer to unseen foundry images with different contrast or noise; testing on SEM images from a different tool or process would clarify the corrector's real-world robustness.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes SEMU-Net, a three-stage pipeline for nanophotonic fabrication correction. A U-Net is first trained to segment SEM images of fabricated silicon photonic structures into silicon/silica masks. A predictor U-Net is then trained to map GDS design files to these segmented SEM images, modeling fabrication-induced variations. Finally, a corrector attention U-Net, trained in tandem with a frozen predictor, maps designs to corrected designs intended to compensate for those variations. The authors report a segmentation IoU of 99.30% and a corrector average IoU of 98.67% on a custom benchmark of stars, gratings, circles, and holes, and claim that the corrected designs ensure the final fabricated structures closely match the intended specifications.

Significance. If the method truly corrected real fabrication variations, it would be a valuable contribution to integrated photonics, where process-induced deviations such as over-etching and corner rounding degrade device performance. The paper has some genuine strengths: it uses real foundry data from Applied Nanotools, compares the segmentation model against SAM and threshold baselines, evaluates several U-Net variants for the corrector, and runs each experiment five times with median reporting. However, the evaluation design prevents the physical claim from being established. The reported corrector IoU is a closed-loop, self-consistency metric computed against the authors' own segmentation and predictor models, with no fabricated corrected devices and no independent measurement of the actual fabricated geometry. The central claim of the abstract is therefore unsupported by the evidence presented.

major comments (4)
  1. [§4.3, Table 2] The headline 98.67% IoU for the corrector is not anchored to a no-correction baseline. To demonstrate that the tandem corrector improves fabrication fidelity, the authors must report the IoU between the original, uncorrected GDS design and the predictor's output (or the segmented SEM) on the same benchmark. Without this baseline, the 98.67% figure is uninterpretable: if the fabrication process already matches the design to roughly 98% IoU, the correction adds negligible value, whereas if the baseline is much lower, the improvement is meaningful. This omission directly affects the quantitative claim made in the abstract.
  2. [§3.2–§3.4] The corrector is trained and evaluated in a closed loop against the authors' own predictor, which in turn is trained on segmentations produced by the authors' own segmentation U-Net (see §3.3, where GDS and SEM images are 'segmented with the help of our segmentation model'). In the tandem architecture, the frozen predictor maps the corrected GDS to a segmented-SEM representation, and the loss penalizes the difference between that prediction and the target segmentation; the reported IoU is therefore a measure of self-consistency within the segmentation/predictor/corrector pipeline, not an agreement with the physical fabricated geometry. The manuscript reports no post-correction SEM images, no optical characterization, and the conclusion explicitly defers fabrication testing to future work, so the abstract's claim that corrected designs 'ensure that the final fabricated structures closely align with the intended specifications' is not supported by the presented evidence.
  3. [§4.3 and §3.4] The ground truth used to compute the corrector IoU is not precisely defined. Section 3.4 states that the tandem loss compares the predictor's output to 'the original SEM image,' while Section 4.3 describes IoU for 'correcting GDS images' without specifying whether the reference is the original GDS, the segmented SEM, or the predictor's output for the corrected design. The paper must state exactly which binary images are compared at evaluation time and how the segmentation model is involved in generating the reference. Without this, the headline 98.67% cannot be reproduced or independently verified.
  4. [§3.3 and §4.1] The size and quality of the manual labels used to train the segmentation model are not reported. The Introduction states that the segmentation model uses 'manually segmented images as labels,' yet Section 3.3 indicates that the SEM images used to train the predictor and corrector are segmented by the segmentation model itself. If the manual label set is small or biased, any systematic segmentation error (e.g., at edges or at silicon/silica boundaries) propagates through the predictor and is then 'corrected' by the corrector, so the high IoU may reflect compensation for segmentation bias rather than compensation for fabrication variations. The authors should report the number of manually labeled images, inter-annotator agreement, and an evaluation of the predictor/corrector on a held-out set with manual ground truth.
minor comments (6)
  1. [§3.2 and Figure 3] The corrector is described in §3.2 as mapping SEM images back to GDS files, but the Figure 3 caption states 'the corrector model takes GDS designs as input and outputs corrected designs.' Please clarify the actual input/output of the corrector and of the tandem configuration.
  2. [§4] The custom benchmark is described as 'several hundred structure images,' but the exact number of images, the train/validation/test split, and the composition of shapes are not given. Please provide these details for reproducibility.
  3. [Tables 1 and 2] The description 'average IoU' is ambiguous; please state whether the IoU is averaged over images or over pixels, and report the standard deviation or interquartile range in addition to the median.
  4. [§4.3] The minimum IoU for the attention U-Net (tandem) is 88.86% while the average is 98.67%; the paper should discuss the cases that produce the low end of the distribution, since a single poorly corrected structure could dominate the practical impact.
  5. [Table 2] The header contains the typo 'Iou'; it should read 'IoU'.
  6. [§3.4] The Dice loss weight of 0.5 in the combined loss is not justified or ablated; please provide a brief ablation or a reference for this choice.

Circularity Check

1 steps flagged · score 6.0 of 10

The corrector's 98.67% IoU is a closed-loop self-consistency score: the corrector is trained and evaluated against a frozen predictor that is itself trained on the authors' segmentation-model outputs, and no corrected layout is fabricated or measured.

  1. fitted input called prediction [Section 3 (Methodology) and Section 3.4 (Tandem Architecture), with results reported in Section 4.3, Table 2]
    "The predictor model specifically evaluates the corrected version of a device image by applying predictions to the corrected design and comparing the resulting output with the original design file. This approach ensures that the corrections are accurately aligned with the intended design specifications."

    In the tandem configuration (Section 3.4), the predictor weights are frozen and the corrector is updated so that the predictor's output on the corrected GDS matches the original design. The reported 98.67% average IoU is measured on this same closed loop: corrected GDS is fed through the same frozen predictor and compared with the target design. The corrector is therefore trained to maximize exactly the quantity that is later reported as the headline accuracy. The 'final fabricated structures' are never measured; they are the predictor's outputs, so the physical claim that corrections make fabricated structures align with specifications reduces to the corrector's ability to invert the individual predictor it was trained with.

full rationale

The paper's central correction claim is not mathematically circular in the sense of an equation reducing to itself, and the corrector does generalize to held-out custom-benchmark shapes. However, the reported 98.67% IoU is a self-consistency metric by construction: Section 3 defines correction quality as the frozen predictor's output on the corrected design matching the original design, and Section 3.4 trains the corrector under that exact objective. Thus the headline accuracy measures P(C(G)) versus G on a benchmark, with P being a learned surrogate rather than a physical fabrication measurement. Because the predictor's training labels are themselves generated by the authors' segmentation U-Net ('The GDS and SEM images are then segmented with the help of our segmentation model,' Section 4.1), any systematic segmentation bias is propagated into the predictor and then 'corrected' for, so the loop is entirely internal to the authors' models. The paper presents no post-correction SEM image, no fabricated corrected layout, and no optical testing, and the conclusion explicitly lists fabrication validation as future work. These features make the physical 'ensuring that the final fabricated structures closely align with the intended specifications' claim unsupported; the IoU score is best understood as a fitting objective rather than an independent evaluation. Self-citations, such as reference [11], are used only to motivate the approach and are not load-bearing, so they do not add to the circularity score. Overall, the core correction 'prediction' reduces to a closed-loop training objective, giving a partial circularity score of 6 rather than a higher score, because the held-out benchmark does provide some generalization evidence for the surrogate-inversion task.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a closed evaluation loop: the segmentation model generates the labels that define the correction target, the predictor learns that mapping, and the corrector inverts the predictor. No independent, measured fabrication data enters the loop, and the proprietary dataset plus undocumented manual labels prevent external verification.

free parameters (3)
  • Dice loss weight = 0.5
    Hand-chosen weight in the combined BCE + 0.5*Dice loss used for predictor and corrector; affects the shape-versus-pixel trade-off, and no sensitivity analysis is provided.
  • Custom benchmark composition = several hundred images of stars, gratings, circles, holes, and more
    The evaluation set is constructed by the authors from the proprietary foundry dataset; size, selection criteria, and difficulty are not specified, so the reported IoU depends on this undocumented choice.
  • Training hyperparameters = lr 0.0001 (seg) / 0.0004 (pred/corr), epochs 50/20/60, patch 256/2048, batch 32/2
    Selected by an unspecified 'thorough search'; no ablations or sensitivity results are shown, so the numbers are not independently checkable.
assumptions (5)
  • domain assumption SEM images faithfully represent the fabricated silicon photonic geometry at 1 nm/pixel resolution.
    The pipeline uses SEM intensity as ground truth for fabrication outcome (Section 4.1). Imaging noise, charging, or calibration errors would appear as fabrication variation.
  • domain assumption Manual labels used to train the segmentation U-Net are accurate enough to serve as ground truth for later models.
    The segmentation model and, through it, the predictor and corrector depend on manual labels whose quality, number, and inter-annotator agreement are not reported (Section 4.2).
  • domain assumption GDS design files and SEM images can be aligned by simple cropping, so pixel-wise comparison is meaningful.
    The pipeline crops and aligns GDS and SEM images before training (Section 4.1). Uncorrected misalignment lowers IoU and is treated as fabrication variation.
  • ad hoc to paper Fabrication variations learned on the custom benchmark shapes (stars, gratings, circles, holes) generalize to real photonic devices.
    The benchmark is designed by the authors to be 'large and varied' with minimal chip space (Section 4.1), but no evidence is given that these test patterns cover the variation modes of real devices.
  • domain assumption The predictor, once trained on GDS-to-segmented-SEM, is a sufficiently accurate model of fabrication that inverting it yields correct compensation.
    The tandem architecture freezes the predictor and optimizes the corrector against it (Section 3.4); if the predictor is inaccurate, the corrector will compensate for the predictor's errors.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SEMU-Net: A Segmentation-based Corrector for Fabrication Process Variations of Nanophotonics with Microscopic Images." pith.science (2026). https://pith.science/paper/VKWQIXZY

@misc{pith2026241116973,
  author       = {Pith},
  title        = {Pith review of: SEMU-Net: A Segmentation-based Corrector for Fabrication Process Variations of Nanophotonics with Microscopic Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VKWQIXZY}},
  note         = {Machine review of arXiv:2411.16973}
}
read the original abstract

Integrated silicon photonic devices, which manipulate light to transmit and process information on a silicon-on-insulator chip, are highly sensitive to structural variations. Minor deviations during nanofabrication-the precise process of building structures at the nanometer scale-such as over- or under-etching, corner rounding, and unintended defects, can significantly impact performance. To address these challenges, we introduce SEMU-Net, a comprehensive set of methods that automatically segments scanning electron microscope images (SEM) and uses them to train two deep neural network models based on U-Net and its variants. The predictor model anticipates fabrication-induced variations, while the corrector model adjusts the design to address these issues, ensuring that the final fabricated structures closely align with the intended specifications. Experimental results show that the segmentation U-Net reaches an average IoU score of 99.30%, while the corrector attention U-Net in a tandem architecture achieves an average IoU score of 98.67%.

Figures

Figures reproduced from arXiv: 2411.16973 by the authors.

Figure 1
Figure 1. Overview of the SEMU-Net framework for improving the fabrication of integrated silicon photonic devices. (i) The segmentation [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. A sample from the ANT NanoSOI dataset: (a) an [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Overview of the segmentation, predictor, and corrector models, all based on the U-Net architecture. The corrector model [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Performance comparison between the attention U-Net in tandem configuration and the original U-Net across three sample [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Workflow of the threshold-based segmentation model. A combination of Canny and Otsu’s methods is used to detect the contours [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Performance comparison between the U-Net, SAM and threshold model for two sample SEM images, which are processed [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 26 canonical work pages

  1. [1]

    Salem Saleh Al-amri, N. V . Kalyankar, and Khamitkar S. D. Image segmentation by using threshold techniques, 2010. 3

  2. [2]

    A review on photonics and its applications

    Kirubel Amsalu and Sivaprakasam Palani. A review on photonics and its applications. Materials Today: Proceed- ings, 33:3372–3377, 2020. International Conference on Nan- otechnology: Ideas, Innovation and Industries. 1

  3. [3]

    Understanding and improving early stopping for learning with noisy labels,

    Yingbin Bai, Erkun Yang, Bo Han, Yanhua Yang, Jiatong Li, Yinian Mao, Gang Niu, and Tongliang Liu. Understanding and improving early stopping for learning with noisy labels,

  4. [4]

    Brown, Benjamin Mann, et al

    Tom B. Brown, Benjamin Mann, et al. Language models are few-shot learners, 2020. 1

  5. [5]

    A computational approach to edge detection

    John Canny. A computational approach to edge detection. IEEE Transactions on Pattern Analysis and Machine Intelli- gence, PAMI-8(6):679–698, 1986. 3

  6. [6]

    Palm: Scaling language modeling with pathways, 2022

    Aakanksha Chowdhery, Sharan Narang, et al. Palm: Scaling language modeling with pathways, 2022. 1

  7. [7]

    When, why and how much? adaptive learning rate scheduling by refinement, 2023

    Aaron Defazio, Ashok Cutkosky, Harsh Mehta, and Kon- stantin Mishchenko. When, why and how much? adaptive learning rate scheduling by refinement, 2023. 5

  8. [8]

    Silicon photonic devices and integrated circuits

    Po Dong, Young-Kai Chen, Guang-Hua Duan, and David Neilson. Silicon photonic devices and integrated circuits. Nanophotonics, 3, 2014. 1

Show all 30 references
  1. [9]

    An image is worth 16x16 words: Transformers for image recognition at scale, 2021

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

  2. [10]

    Gatys, Alexander S

    Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge. Image style transfer using convolutional neural networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 2414–2423, 2016. 3

  3. [11]

    Deep learning-based prediction of fabrication-process-induced structural variations in nanophotonic devices

    Dusan Gostimirovic, Dan-Xia Xu, Odile Liboiron- Ladouceur, and Yuri Grinberg. Deep learning-based prediction of fabrication-process-induced structural variations in nanophotonic devices. ACS Photonics , 9(8):2623–2633, 2022. 1, 3

  4. [12]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-image translation with conditional adver- sarial networks, 2018. 3

  5. [13]

    A survey of loss functions for semantic seg- mentation

    Shruti Jadon. A survey of loss functions for semantic seg- mentation. In 2020 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB), pages 1–7. IEEE, Oct 2020. 5

  6. [14]

    Berg, Wan-Yen Lo, Piotr Doll ´ar, and Ross Girshick

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C. Berg, Wan-Yen Lo, Piotr Doll ´ar, and Ross Girshick. Segment anything, 2023. 6

  7. [15]

    Per- formance prediction for silicon photonics integrated circuits with layout-dependent correlated manufacturing variability

    Zeqin Lu, Jaspreet Jhoja, Jackson Klein, Xu Wang, Amy Liu, Jonas Flueckiger, James Pond, and Lukas Chrostowski. Per- formance prediction for silicon photonics integrated circuits with layout-dependent correlated manufacturing variability. Optics Express, 25(9):9712–9733, 2017. 3

  8. [16]

    Raunet: Residual attention u-net for semantic segmentation of cataract surgical instruments, 2019

    Zhen-Liang Ni, Gui-Bin Bian, Xiao-Hu Zhou, Zeng-Guang Hou, Xiao-Liang Xie, Chen Wang, Yan-Jie Zhou, Rui-Qi Li, and Zhen Li. Raunet: Residual attention u-net for semantic segmentation of cataract surgical instruments, 2019. 2, 7

  9. [17]

    Hlaing, David Z

    Shupeng Ning, Hanqing Zhu, Chenghao Feng, Jiaqi Gu, Zhixing Jiang, Zhoufeng Ying, Jason Midkiff, Sourabh Jain, May H. Hlaing, David Z. Pan, and Ray T. Chen. Photonic- electronic integrated circuits for high-performance comput- ing and ai accelerators, 2024. 1

  10. [18]

    Hammerla, Bernhard Kainz, Ben Glocker, and Daniel Rueckert

    Ozan Oktay, Jo Schlemper, Loic Le Folgoc, Matthew Lee, Mattias Heinrich, Kazunari Misawa, Kensaku Mori, Steven McDonagh, Nils Y . Hammerla, Bernhard Kainz, Ben Glocker, and Daniel Rueckert. Attention u-net: Learning where to look for the pancreas, 2018. 2, 3, 4, 7

  11. [19]

    An introduction to convolu- tional neural networks, 2015

    Keiron O’Shea and Ryan Nash. An introduction to convolu- tional neural networks, 2015. 1

  12. [20]

    Piggott, Eric Y

    Alexander Y . Piggott, Eric Y . Ma, Logan Su, Geun Ho Ahn, Neil V . Sapra, Dries Vercruysse, Andrew M. Nether- ton, Akhilesh S. P. Khope, John E. Bowers, and Jelena Vuˇckovi´c. Inverse-designed photonics for semiconductor foundries. ACS Photonics, 7(3):569–575, 2020. 1

  13. [21]

    U-net: Convolutional networks for biomedical image segmentation,

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation,

  14. [22]

    Schubert, Alfred K

    Martin F. Schubert, Alfred K. C. Cheung, Ian A. D. Williamson, Aleksandra Spyra, and David H. Alexander. In- verse design of photonic devices with strict foundry fabrica- tion constraints. ACS Photonics, 9(7):2327–2336, 2022. 2

  15. [23]

    Shawn Yohanes Siew, Bo Li, Feng Gao, H. Y . Zheng, Weipeng Zhang, Pengfei Guo, S. W. Xie, A. Song, Bo Dong, L. W. Luo, Chao Li, Xianshu Luo, and GuoQiang Lo. Re- view of silicon photonics technology and platform develop- ment. Journal of Lightwave Technology, 39(13):4374–4389,

  16. [24]

    A Re- view of Object Detection Models Based on Convolutional Neural Network, pages 1–16

    Farhana Sultana, Abu Sufian, and Paramartha Dutta. A Re- view of Object Detection Models Based on Convolutional Neural Network, pages 1–16. Springer Singapore, 2020. 1

  17. [25]

    Deep image: Scaling up image recognition, 2015

    Ren Wu, Shengen Yan, Yi Shan, Qingqing Dang, and Gang Sun. Deep image: Scaling up image recognition, 2015. 1

  18. [26]

    Multi-scale context aggrega- tion by dilated convolutions, 2016

    Fisher Yu and Vladlen Koltun. Multi-scale context aggrega- tion by dilated convolutions, 2016. 3

  19. [27]

    Hyper-parameter optimization: A review of algorithms and applications, 2020

    Tong Yu and Hong Zhu. Hyper-parameter optimization: A review of algorithms and applications, 2020. 5

  20. [28]

    Unet++: A nested u-net architecture for medical image segmentation, 2018

    Zongwei Zhou, Md Mahfuzur Rahman Siddiquee, Nima Tajbakhsh, and Jianming Liang. Unet++: A nested u-net architecture for medical image segmentation, 2018. 2, 7

  21. [29]

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. Unpaired image-to-image translation using cycle- consistent adversarial networks, 2020. 3

  22. [30]

    Manmatha, and Mu Li

    Yi Zhu, Xinyu Li, Chunhui Liu, Mohammadreza Zolfaghari, Yuanjun Xiong, Chongruo Wu, Zhi Zhang, Joseph Tighe, R. Manmatha, and Mu Li. A comprehensive study of deep video action recognition, 2020. 1

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.