Pith. sign in

REVIEW 3 major objections 5 minor 37 references

GasTwinFormer: A Hybrid Vision Transformer for Livestock Methane Emission Segmentation and Dietary Classification in Optical Gas Imaging

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A 3.3-million-parameter model segments methane plumes in cattle barns and reads the animals' diet from the same video.

desk verdict The segmentation work and new dataset are worth a look, but the 100% diet classification is a temporal-split artifact and the test-set selection and annotation pipeline need fixing before the claims can stand. read the letter →

arxiv 2508.15057 v1 pith:UJXAB5N7 submitted 2025-08-20 cs.CV

Toqi Tahamid Sarker , Mohamed Embaby , Taminul Islam , Amer AbuGhazaleh , Khaled R Ahmed This is my paper · ORCID
classification cs.CV
keywords livestockmethaneemissionsopticalgasimagingsemanticsegmentationvisiontransformerhybridattentiondietaryclassificationplumereal-timemonitoring
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

Livestock methane is a large slice of human-caused methane, and today's monitoring options are either costly chambers or offline analysis of gas-imaging footage. GasTwinFormer is designed to show that one compact hybrid vision transformer can do both jobs a barn monitor needs: pixel-level segmentation of methane plumes in optical gas images and scene-level classification of the animal's dietary treatment. The paper contributes the first beef-cattle optical-gas-imaging benchmark, 11,694 annotated frames from 19 videos across high-forage, mixed, and high-grain diets, and reports 74.47% mIoU and 83.63% mF1 for segmentation at 3.348M parameters and 114.9 FPS, plus perfect dietary classification. If the results hold, a single small model could watch eructation plumes continuously and report both where the methane is and which feeding regimen produced it.

What carries the argument

Mix Twin encoder: four hierarchical stages, each pairing an Efficient Multi-Head Attention block (EMA) with a Locally-Grouped Self-Attention block (LSA). EMA reduces the key/value spatial resolution by a stage-dependent factor R to make global attention affordable; LSA splits the feature map into non-overlapping 5×5 windows and runs attention inside each, recovering local boundary detail. LR-ASPP decoder: the deepest feature goes through a lightweight atrous spatial pyramid pooling path while the three shallower levels are fused progressively through 1×1 convolutions, so plume boundaries keep their fine spatial resolution. Gaussian Plume Weighted Dice Loss: weights each pixel by a Gaussian c

What would settle it

Hold out entire cows or entire recording sessions when splitting train and test for the dietary classifier; if accuracy collapses toward chance on unseen cows while staying 100% on seen cows, the diet-emission claim is refuted. A second check: shuffle diet labels within each session and retrain — a classifier that still scores 100% cannot be learning diet.

Watch

Extended reading notes

Core claim

The central claim is that alternating a spatially-reduced global attention block with a locally-grouped attention block in every encoder stage creates a better backbone for methane plume segmentation than either attention style alone. The GasTwinFormer encoder stacks four stages, each containing one EMA block (keys and values spatially reduced to cut attention cost) followed by one LSA block (5×5 local windows), then feeds the four feature levels into a hierarchical LR-ASPP decoder that fuses shallow details with deep semantics. A Gaussian Plume Weighted Dice Loss nudges the segmentation toward the expected bell-shaped diffusion of a plume, and a small head on the deepest features predicts t

Load-bearing premise

The perfect dietary-classification score assumes the model is reading diet-driven differences in the methane plumes, not cues tied to which cow, barn, or recording session each frame came from; with only four cows per diet and time-contiguous train/test splits, the classification head could be memorizing video identity instead.

Editorial extensions

If this is right

  • Continuous real-time plume monitoring is feasible on modest hardware, because the model runs at 114.9 FPS with 3.348M parameters and 3.428 GFLOPs.
  • Dietary-treatment effects could be monitored passively from OGI footage without separate diet records or sensors, because the classification head reads the treatment directly from the plume images.
  • The design recipe — EMA-then-LSA stages, LR-ASPP fusion, and Gaussian plume weighting — transfers to other gas-imaging segmentation tasks, since ablations show each component contributes.
  • The 11,694-frame, three-diet benchmark gives later work a common test set for livestock methane segmentation and dietary classification.
  • The reported gains over pure-global, pure-local, and CNN baselines support the broader claim that hybrid global-local attention is worth its small computational cost in low-signal thermal imagery.

Reading between the lines

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

  • An immediate next test is to split the diet classifier by cow or by recording session: only then can the 100% accuracy be attributed to diet rather than to barn, lighting, or animal identity, since the current splits are time-contiguous frames from the same 19 videos and 12 cows.
  • The Gaussian-plume weighting suggests a natural extension from 'where is the plume' to 'how much methane': the same masks and plume statistics could be calibrated against controlled-release measurements to estimate emission rates.
  • The architecture is not cattle-specific; transferring the encoder/decoder to industrial leak detection or other ruminant species is a plausible next step, with window size and loss parameters retuned.
  • Since only plume frames are annotated and non-plume frames are discarded, a practical barn monitor would still need a trigger to start segmentation; pairing GasTwinFormer with a cheap frame-level plume detector is an open engineering question the paper leaves implicit.
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

3 major / 5 minor

Summary. GasTwinFormer proposes a hybrid transformer encoder (EMA + LSA, termed Mix Twin), an LR-ASPP decoder, and a dietary classification head for joint methane plume segmentation and diet classification in optical gas imaging. The paper also introduces a semi-automatically annotated beef cattle OGI dataset of 11,694 frames from 19 videos across three diets. The main claims are 74.47% mIoU / 83.63% mF1 segmentation, 100% dietary classification accuracy, and efficient real-time deployment (3.348M parameters, 3.428G FLOPs, 114.9 FPS). Extensive ablations compare decoder heads, fusion branches, attention patterns, loss functions, and LSA window sizes.

Significance. If the reported numbers are reliable, the architecture offers a strong efficiency-accuracy trade-off for real-time OGI segmentation, and the dataset would be a useful resource for livestock methane monitoring. The paper provides a systematic architectural study and a broad comparison against transformer and CNN baselines, which are strengths. However, the two headline empirical claims—superior segmentation and, in particular, perfect diet classification—rest on methodological choices that are not currently convincing: temporal split leakage, annotation-pipeline circularity, and model selection on the test set. The abstract's conclusion about 'leveraging diet-emission correlations' is especially vulnerable because the classification experiment cannot distinguish diet from cow/barn/session cues.

major comments (3)
  1. [Sec. 4, Table 1; Sec. 5.2, Table 2] The dietary classification evaluation uses a temporal split inside each video, so frames from the same cows and the same recording sessions appear in both training and test. With only 12 cows (4 per diet), each group housed in separate feed stalls, the reported 100% accuracy can be explained by cow identity, stall background, lighting, or recording session rather than by diet-specific emission patterns. The fact that nearly every strong baseline also reaches 100% in Table 2 strongly supports a trivial cue. Please re-evaluate with leave-one-cow-out or leave-one-video-out splits and report per-cow accuracy. This is load-bearing for the central 'leveraging diet-emission correlations' claim.
  2. [Sec. 4, Annotation Methodology; Sec. 5.2, Table 2] Ground-truth masks are generated by a semi-automatic pipeline in which Gasformer (ref [19]) trained on classical masks produces candidate masks, and manual inspection selects among three candidates. Gasformer is also a primary segmentation baseline in Table 2. This makes the benchmark circular with respect to the transformer approach and can inflate the reported gains over CNN baselines and the absolute mIoU. Please provide an independently human-annotated test subset, report inter-annotator agreement, and show that model rankings are stable on that subset.
  3. [Sec. 5.1; Sec. 5.3, Tables 3–5] All metrics are computed on the test set, and the final configuration (EL-EL-EL-EL, 5x5 window, Gaussian Plume loss) was chosen by comparing test mIoU across a large set of ablations. Differences between near-tied configurations are only 0.04 (EL vs LE), 0.12 (5x5 vs 3x3), and 0.28 (LR-ASPP vs all-MLP) mIoU, with no repeated runs or error bars. This constitutes model selection on the test set and undermines the precision of the headline 74.47% mIoU. Use the validation set for model selection and report confidence intervals over multiple seeds.
minor comments (5)
  1. [Sec. 1 vs Sec. 4] The introduction states that OGI cameras operate in the 7–8.5 µm spectral range, while Sec. 4 specifies the FLIR Gx320 camera as 3.2–3.4 µm. The latter is consistent with methane absorption; please correct the introduction.
  2. [Sec. 3.1] The text says the spatial reduction process is 'introduced in Pyramid Vision Transformer [30]', but the PVT reference is [28]. Please fix the citation.
  3. [Sec. 3.4, Eq. (6)] The plume center (µx, µy) is computed via center-of-mass on predicted masks. This makes the weighting prediction-dependent and potentially unstable when predictions are poor. Please clarify whether this is used at training and/or inference and whether the reported gain is robust to this choice.
  4. [Table 2] The caption says 'Bold indicates better', but bold formatting in the table is inconsistent across columns and rows. Please make the best value in each column explicit.
  5. [Sec. 4, Dataset Statistics] The camera sensor is 320x240 but frames are stored as 640x480 PNGs. Please clarify whether the frames were upsampled from the native resolution and describe the resampling procedure.

Circularity Check

2 steps flagged · score 6.0 of 10

Benchmark ground truth is partly generated by the authors' own Gasformer model, and the perfect diet classification is confounded by within-video temporal splits.

  1. self citation load bearing [Section 4, Annotation Methodology; Section 5.2, Table 2]
    "(2) Deep learning processing uses a Gasformer [19] model trained on initial classical masks to identify subtle patterns beyond traditional methods. ... For each frame, we generate three mask candidates from these approaches and perform manual inspection to select the most accurate representation."

    The ground-truth masks of the new benchmark are partly produced by Gasformer (ref [19]), which is also the main transformer baseline in Table 2. Thus the mIoU comparisons—including the headline 2.22% gain over Gasformer—are evaluated against a label set generated with Gasformer's own inductive biases. This is a self-referential benchmark: the proposed model inherits SegFormer-style components from the same family, so the labels are not an independent measurement. The manual inspection step does not break the dependency because the candidate set is already narrowed toward the classical/Gasformer pipeline.

  2. fitted input called prediction [Section 4, Dataset Statistics and Composition; Section 5.2, Dietary Classification Performance; Abstract]
    "For model development, we employed temporal splitting ... 70% of consecutive frames for training, 15% for validation, and 15% for testing within each video. ... All 19 videos contribute to each split while maintaining dietary treatment proportions. ... Additionally, our method achieves perfect dietary classification accuracy (100%), demonstrating the effectiveness of leveraging diet-emission correlations."

    The dietary-label prediction is forced by the split: every video contributes frames to both training and test, so the same cows, chutes, lighting, and barn backgrounds appear on both sides. Since diet label is constant per video, a model can memorize video-level cues and hit 100% without reading methane plumes; Table 2 confirms nearly all strong baselines also achieve 100%. The claimed demonstration of diet-emission correlations is therefore an artifact of the experimental design, not an independent prediction.

full rationale

The paper's segmentation architecture is not itself circular: the Mix Twin encoder combines published EMA and LSA mechanisms, and the ablations (decoder head, fusion, attention pattern, loss, window size) are self-contained experiments against a fixed label set. The circularity enters through the construction of that label set and through the dietary-classification evaluation. First, the ground-truth masks are 'semi-automatically' produced with the authors' own Gasformer (ref [19]) as one of three candidates, and Gasformer is the paper's key baseline; the evaluation therefore contains a self-citation loop (labels generated by the comparison model). This contaminates all mIoU numbers, including the relative ranking, even though the manual inspection and classical candidates give it partial independent content. Second, the claimed 100% dietary classification is not an out-of-sample test of diet-emission correlations: the temporal split keeps all 19 videos in train, validation, and test, so the same cow and barn appear in both training and test. With only 12 cows (4 per diet) in separate stalls, the model can trivially identify the video/stall. The near-universal 100% accuracy across all strong baselines confirms a trivial cue. These two issues together warrant a partial-circularity score of 6; the architecture and ablation content prevent a higher score.

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

The load-bearing result is an empirical comparison on a new dataset. The largest implicit burdens sit in the labels (hand-tuned classical thresholds plus the authors' own Gasformer as a candidate generator) and in the interpretation of the 100% classification score (diet as the dominant, confound-free signal). No new physical entities are postulated.

free parameters (5)
  • LSA window size = 5x5 at 512x512 input
    Selected by comparing test mIoU across 3x3, 5x5, and 7x7 windows (Sec. 5.3, Table 5).
  • Gaussian plume diffusion bounds (sigma_x, sigma_y) = clamped to [W/20, W/2] and [H/20, H/2]
    Hand-set adaptive bounds for the weighted Dice loss; controls spatial weighting of the loss (Sec. 3.4).
  • Annotation pipeline thresholds = motion 20-60; adaptive block 300-500; constant 5-15; min size 2000 px; eccentricity >0.95; CLAHE; h=15
    Hand-chosen parameters of the semi-automatic ground-truth pipeline; they directly shape the benchmark labels (Sec. 4).
  • Decoder channels and classification head width = 128 LR-ASPP channels; 256 hidden channels
    Selected by ablation over 128-2048 channel counts and feature-source stages (Sec. 5.3, Table 4).
  • Stage-adaptive attention reduction ratios R = {8, 4, 2, 1}
    Adopted from SegFormer/PVT design; not optimized here, but per-stage and part of the architecture's complexity profile (Sec. 3.1).
assumptions (5)
  • domain assumption Methane plumes are visually segmentable in 3.2-3.4 um OGI grayscale frames, and the 11,694 selected frames are representative of emission events.
    Underlies the dataset construction and every segmentation metric (Sec. 4).
  • ad hoc to paper Ground-truth masks from the semi-automatic pipeline, including Gasformer-generated candidates, are accurate enough to rank models.
    The authors' own model produced candidate masks; no inter-annotator agreement or independent sensor validation is reported (Sec. 4).
  • ad hoc to paper Differences across dietary groups dominate other frame-level differences (cow identity, barn, lighting) that the classifier could exploit.
    Required for 100% accuracy to mean diet-emission correlation rather than video identity (Sec. 5.2, Table 2).
  • domain assumption ImageNet-pretrained weights transfer to thermal grayscale OGI imagery.
    All compared models use ImageNet initialization (Sec. 5.1).
  • domain assumption The Gaussian plume model is a suitable spatial prior for eructation plumes, and its center computed from predicted masks is a stable reference.
    Motivates the weighted Dice loss; the plume center is taken from the prediction itself, which makes the weights prediction-dependent (Sec. 3.4).

how reviews work

0 comments
Cite this review

Pith. "Pith review of GasTwinFormer: A Hybrid Vision Transformer for Livestock Methane Emission Segmentation and Dietary Classification in Optical Gas Imaging." pith.science (2026). https://pith.science/paper/UJXAB5N7

@misc{pith2026250815057,
  author       = {Pith},
  title        = {Pith review of: GasTwinFormer: A Hybrid Vision Transformer for Livestock Methane Emission Segmentation and Dietary Classification in Optical Gas Imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UJXAB5N7}},
  note         = {Machine review of arXiv:2508.15057}
}
read the original abstract

Livestock methane emissions represent 32% of human-caused methane production, making automated monitoring critical for climate mitigation strategies. We introduce GasTwinFormer, a hybrid vision transformer for real-time methane emission segmentation and dietary classification in optical gas imaging through a novel Mix Twin encoder alternating between spatially-reduced global attention and locally-grouped attention mechanisms. Our architecture incorporates a lightweight LR-ASPP decoder for multi-scale feature aggregation and enables simultaneous methane segmentation and dietary classification in a unified framework. We contribute the first comprehensive beef cattle methane emission dataset using OGI, containing 11,694 annotated frames across three dietary treatments. GasTwinFormer achieves 74.47% mIoU and 83.63% mF1 for segmentation while maintaining exceptional efficiency with only 3.348M parameters, 3.428G FLOPs, and 114.9 FPS inference speed. Additionally, our method achieves perfect dietary classification accuracy (100%), demonstrating the effectiveness of leveraging diet-emission correlations. Extensive ablation studies validate each architectural component, establishing GasTwinFormer as a practical solution for real-time livestock emission monitoring. Please see our project page at gastwinformer.github.io.

Figures

Figures reproduced from arXiv: 2508.15057 by the authors.

Figure 1
Figure 1. GasTwinFormer architecture. The Mix Twin encoder uses alternating EMA and LSA blocks across four hierarchical stages. The [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Qualitative comparison of methane plume segmentation results across different models and dietary treatments (MD: mixed diet, [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 30 canonical work pages

  1. [19]

    Gasformer: A transformer-based architecture for segmenting methane emissions from livestock in optical gas imaging

    Toqi Tahamid Sarker, Mohamed G Embaby, Khaled R Ahmed, and Amer AbuGhazaleh. Gasformer: A transformer-based architecture for segmenting methane emissions from livestock in optical gas imaging. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5489–5497, 2024. 2, 5, 6

    work page 2024

  2. [1]

    Rethinking atrous convolution for seman- tic image segmentation

    Liang-Chieh Chen, George Papandreou, Florian Schroff, and Hartwig Adam. Rethinking atrous convolution for seman- tic image segmentation. arXiv preprint arXiv:1706.05587 ,

    arXiv

  3. [2]

    Twins: Revisiting the design of spatial attention in vision transformers

    Xiangxiang Chu, Zhi Tian, Yuqing Wang, Bo Zhang, Haib- ing Ren, Xiaolin Wei, Huaxia Xia, and Chunhua Shen. Twins: Revisiting the design of spatial attention in vision transformers. Advances in neural information processing systems, 34:9355–9366, 2021. 1, 2, 3, 5, 6

    work page 2021

  4. [3]

    An image is worth 16x16 words: Trans- formers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, et al. An image is worth 16x16 words: Trans- formers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. 2

    arXiv 2010

  5. [4]

    Optical gas imaging and deep learning for quantifying enteric methane emissions from rumen fermentation in vitro

    Mohamed G Embaby, Toqi Tahamid Sarker, Amer AbuG- hazaleh, and Khaled R Ahmed. Optical gas imaging and deep learning for quantifying enteric methane emissions from rumen fermentation in vitro. IET Image Processing , 19(1):e13327, 2025. 1

    work page 2025

  6. [5]

    Searching for mo- bilenetv3

    Andrew Howard, Mark Sandler, Grace Chu, Liang-Chieh Chen, Bo Chen, Mingxing Tan, Weijun Wang, Yukun Zhu, Ruoming Pang, Vijay Vasudevan, et al. Searching for mo- bilenetv3. In Proceedings of the IEEE/CVF international conference on computer vision, pages 1314–1324, 2019. 1, 3, 4, 8

    work page 2019

  7. [6]

    Interlaced sparse self-attention for semantic segmentation

    Lang Huang, Yuhui Yuan, Jianyuan Guo, Chao Zhang, Xilin Chen, and Jingdong Wang. Interlaced sparse self-attention for semantic segmentation. arXiv preprint arXiv:1907.12273, 2019. 1, 7, 8

    arXiv 1907

  8. [7]

    Deep learning-based quantifications of methane emissions with field applications

    Ismot Jahan, Mohamed Mehana, Georgios Matheou, and Hari Viswanathan. Deep learning-based quantifications of methane emissions with field applications. International Journal of Applied Earth Observation and Geoinformation , 132:104018, 2024. 1, 2

    work page 2024

Show all 37 references
  1. [8]

    Local-to-global self-attention in vision trans- formers

    Jinpeng Li, Yichao Yan, Shengcai Liao, Xiaokang Yang, and Ling Shao. Local-to-global self-attention in vision trans- formers. arXiv preprint arXiv:2107.04735, 2021. 1

    2021 arXiv

  2. [9]

    Expediting large-scale vision transformer for dense predic- tion without fine-tuning

    Weicong Liang, Yuhui Yuan, Henghui Ding, Xiao Luo, Wei- hong Lin, Ding Jia, Zheng Zhang, Chao Zhang, and Han Hu. Expediting large-scale vision transformer for dense predic- tion without fine-tuning. Advances in Neural Information Processing Systems, 35:35462–35477, 2022. 1

    2022

  3. [10]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, pages 10012–10022, 2021. 2

    2021

  4. [11]

    Fully convolutional networks for semantic segmentation

    Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In Pro- ceedings of the IEEE conference on computer vision and pat- tern recognition, pages 3431–3440, 2015. 8

    2015

  5. [12]

    Institutional animal care and use committee guidebook

    Office of Laboratory Animal Welfare. Institutional animal care and use committee guidebook. Guidebook, National Institutes of Health, Bethesda, MD, 2002. 5

    2002

  6. [13]

    Advancements in real-time monitoring of en- teric methane emissions from ruminants

    Se ´an O’Connor, Flannag ´an Noonan, Desmond Savage, and Joseph Walsh. Advancements in real-time monitoring of en- teric methane emissions from ruminants. Agriculture, 14(7): 1096, 2024. 1

    2024

  7. [14]

    Cli- mate change 2014: synthesis report

    Rajendra K Pachauri, Myles R Allen, Vicente R Barros, John Broome, Wolfgang Cramer, Renate Christ, John A Church, Leon Clarke, Qin Dahe, Purnamita Dasgupta, et al. Cli- mate change 2014: synthesis report. Contribution of Work- ing Groups I, II and III to the fifth assessment r...

    2014

  8. [15]

    Deep dual-resolution networks for real-time and accurate se- mantic segmentation of traffic scenes

    Huihui Pan, Yuanduo Hong, Weichao Sun, and Yisong Jia. Deep dual-resolution networks for real-time and accurate se- mantic segmentation of traffic scenes. IEEE Transactions on Intelligent Transportation Systems, 2022. 5, 6

    2022

  9. [16]

    Fast-scnn: Fast semantic segmentation network

    Rudra PK Poudel, Stephan Liwicki, and Roberto Cipolla. Fast-scnn: Fast semantic segmentation network. arXiv preprint arXiv:1902.04502, 2019. 5, 6

    1902 arXiv

  10. [17]

    Vi- sion transformers for dense prediction

    Ren ´e Ranftl, Alexey Bochkovskiy, and Vladlen Koltun. Vi- sion transformers for dense prediction. In Proceedings of the IEEE/CVF international conference on computer vision, pages 12179–12188, 2021. 2

    2021

  11. [18]

    Automatic de- tection of methane emissions in multispectral satellite im- agery using a vision transformer

    Bertrand Rouet-Leduc and Claudia Hulbert. Automatic de- tection of methane emissions in multispectral satellite im- agery using a vision transformer. Nature Communications, 15(1):3801, 2024. 2

    2024

  12. [20]

    Creating a sustainable food future: a menu of solutions to feed nearly 10 billion people by 2050-synthesis report

    Tim Searchinger, Richard Waite, Craig Hanson, Janet Ran- ganathan, Patrice Dumas, and Emily Matthews. Creating a sustainable food future: a menu of solutions to feed nearly 10 billion people by 2050-synthesis report. 2018. 1

  13. [21]

    Global methane assessment: Benefits and costs of mitigating methane emissions

    Drew Shindell, AR Ravishankara, Johan CI Kuylenstierna, Eleni Michalopoulou, Lena H ¨oglund-Isaksson, Yuqiang Zhang, Karl Seltzer, Muye Ru, Rithik Castelino, Greg Falu- vegi, et al. Global methane assessment: Benefits and costs of mitigating methane emissions. Technical report...

    2021

  14. [22]

    Quantification of methane emitted by ruminants: a review of methods

    Luis Orlindo Tedeschi, Adibe Luiz Abdalla, Clementina Al- varez, Samuel Weniga Anuga, Jacobo Arango, Karen A Beauchemin, Philippe Becquet, Alexandre Berndt, Robert Burns, Camillo De Camillis, et al. Quantification of methane emitted by ruminants: a review of methods. Journal o...

    2022

  15. [23]

    FLIR G-Series: Gx320, G620, Gx620 Op- tical Gas Imaging (OGI) Cameras for Hydrocarbons

    Teledyne FLIR. FLIR G-Series: Gx320, G620, Gx620 Op- tical Gas Imaging (OGI) Cameras for Hydrocarbons . Tele- dyne FLIR, LLC, 2023. Datasheet. 4

    2023

  16. [24]

    Repvit: Revisiting mobile cnn from vit perspective

    Ao Wang, Hui Chen, Zijia Lin, Jungong Han, and Guiguang Ding. Repvit: Revisiting mobile cnn from vit perspective. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15909–15920, 2024. 5, 6

    2024

  17. [25]

    Machine vision for natural gas methane emissions detection using an infrared camera

    Jingfan Wang, Lyne P Tchapmi, Arvind P Ravikumar, Mike McGuire, Clay S Bell, Daniel Zimmerle, Silvio Savarese, 9 and Adam R Brandt. Machine vision for natural gas methane emissions detection using an infrared camera. Applied En- ergy, 257:113998, 2020. 1, 2

    2020

  18. [26]

    Videogasnet: Deep learning for natural gas methane leak classification using an infrared camera

    Jingfan Wang, Jingwei Ji, Arvind P Ravikumar, Silvio Savarese, and Adam R Brandt. Videogasnet: Deep learning for natural gas methane leak classification using an infrared camera. Energy, 238:121516, 2022. 2

    2022

  19. [27]

    Large- scale controlled experiment demonstrates effectiveness of methane leak detection and repair programs at oil and gas fa- cilities

    Jiayang Lyra Wang, Brenna Barlow, Wes Funk, Cooper Robinson, Adam Brandt, and Arvind P Ravikumar. Large- scale controlled experiment demonstrates effectiveness of methane leak detection and repair programs at oil and gas fa- cilities. Environmental Science & Technology, 58(7):...

    2024

  20. [28]

    Pyramid vision transformer: A versatile backbone for dense prediction without convolutions

    Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and Ling Shao. Pyramid vision transformer: A versatile backbone for dense prediction without convolutions. In Proceedings of the IEEE/CVF international conference on computer vision , p...

    2021

  21. [29]

    Unified perceptual parsing for scene understand- ing

    Tete Xiao, Yingcheng Liu, Bolei Zhou, Yuning Jiang, and Jian Sun. Unified perceptual parsing for scene understand- ing. In Proceedings of the European Conference on Com- puter Vision (ECCV), pages 418–434, 2018. 5, 6

    2018

  22. [30]

    Segformer: Simple and efficient design for semantic segmentation with transform- ers

    Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M Alvarez, and Ping Luo. Segformer: Simple and efficient design for semantic segmentation with transform- ers. Advances in neural information processing systems, 34: 12077–12090, 2021. 1, 2, 3, 4, 5, 6, 8

    2021

  23. [31]

    Transformer-based attention networks for continuous pixel-wise prediction

    Guanglei Yang, Hao Tang, Mingli Ding, Nicu Sebe, and Elisa Ricci. Transformer-based attention networks for continuous pixel-wise prediction. In Proceedings of the IEEE/CVF International Conference on Computer vision , pages 16269–16279, 2021. 1, 2

    2021

  24. [32]

    Bisenet: Bilateral segmentation network for real-time semantic segmentation

    Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, and Nong Sang. Bisenet: Bilateral segmentation network for real-time semantic segmentation. In Proceed- ings of the European conference on computer vision (ECCV), pages 325–341, 2018. 5, 6

    2018

  25. [33]

    Bisenet v2: Bilateral net- work with guided aggregation for real-time semantic seg- mentation

    Changqian Yu, Changxin Gao, Jingbo Wang, Gang Yu, Chunhua Shen, and Nong Sang. Bisenet v2: Bilateral net- work with guided aggregation for real-time semantic seg- mentation. International Journal of Computer Vision, pages 1–18, 2021. 5, 6

    2021

  26. [34]

    Icnet for real-time semantic segmenta- tion on high-resolution images

    Hengshuang Zhao, Xiaojuan Qi, Xiaoyong Shen, Jianping Shi, and Jiaya Jia. Icnet for real-time semantic segmenta- tion on high-resolution images. In Proceedings of the Eu- ropean conference on computer vision (ECCV) , pages 405– 420, 2018. 5, 6

    2018

  27. [35]

    iformer: Integrating convnet and transformer for mobile application

    Chuanyang Zheng. iformer: Integrating convnet and transformer for mobile application. arXiv preprint arXiv:2501.15369, 2025. 5, 6

    2025 arXiv

  28. [36]

    High-accuracy combustible gas cloud imaging system using yolo-plume classification net- work

    Jiani Zhou, Yang Liu, Yong Zhang, Haotian Hu, Zenan Leng, Feng Sun, and Chen Chen. High-accuracy combustible gas cloud imaging system using yolo-plume classification net- work. Frontiers in Physics, 13:1603047, 2025. 4, 8

    2025

  29. [37]

    Asymmetric non-local neural networks for semantic segmentation

    Zhen Zhu, Mengde Xu, Song Bai, Tengteng Huang, and Xi- ang Bai. Asymmetric non-local neural networks for semantic segmentation. In Proceedings of the IEEE/CVF international conference on computer vision, pages 593–602, 2019. 7, 8 10

    2019

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.