Pith. sign in

REVIEW 3 major objections 7 minor 43 references

DISCo: Deep learning, Instance Segmentation, and Correlations for cell segmentation in calcium imaging

T0 review · 3 major / 7 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper claims that DISCo, by combining segment-wise pixel correlations with a summary image and partitioning a signed graph of learned affinities, outperforms all other methods trained on the Neurofinder calcium-imaging datasets.

desk verdict Solid methods paper with a real pipeline advance, but the 'outperforms all' claim is slightly ahead of the evidence. read the letter →

arxiv 1908.07957 v4 pith:ANFE5W6T submitted 2019-08-21 q-bio.NC cs.LGeess.IV

Elke Kirschbaum , Alberto Bailoni , Fred A. Hamprecht This is my paper · ORCID
classification q-bio.NCcs.LGeess.IV
keywords calciumimagingcellsegmentationinstancepixelaffinitiescorrelationclusteringdeeplearningsignedgraphpartitioningNeurofinderbenchmark
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 aims to establish that a cell-segmentation method for calcium-imaging videos can outperform existing models on the public Neurofinder benchmark by combining two complementary cues: temporal correlations between pixels and a static shape-based summary image. Rather than classifying foreground pixels, DISCo predicts pairwise affinities between nearby pixels and extracts cells by partitioning a signed graph built from those affinities. The main numerical claim is that per-dataset-series training (DISCos) achieves the highest average F1 (0.67) among methods trained only on the Neurofinder training data, slightly ahead of the 3dCNN baseline (0.66). Lesion studies support the core design: neither correlations alone (0.50) nor summary images alone (0.49) match the combination, and segment-wise correlations with a learned aggregator beat whole-video correlations and fixed statistics.

What carries the argument

The load-bearing object is the predicted pixel-affinity field combined with the GASP signed-graph partitioner. Affinities $a_{ij}\in[0,1]$ are converted to edge weights $w_{ij}=a_{ij}-0.5$, so positive weights attract pixels into the same cluster and negative weights separate them; GASP then merges clusters while average linkage is positive. This turns instance segmentation into correlation clustering, solved approximately by a greedy agglomerative algorithm. The other essential component is the input representation: segment-wise Pearson correlations ($N=10$ segments, $C=15$ neighbor offsets) aggregated by a 3D CNN, concatenated with a mean-intensity projection.

What would settle it

Rerun the released DISCo code on the Neurofinder training set, train per-series models with several random seeds, and submit to the Neurofinder test set; the claim fails if the average F1 does not reliably stay above 0.66, the score reported for 3dCNN.

Watch

Extended reading notes

Core claim

DISCo treats cell segmentation in calcium-imaging videos as a graph-partitioning problem on learned pixel affinities. The video is split into ten temporal segments; for each pixel, Pearson correlations are computed against fifteen neighboring pixels in each segment, producing a stack that a small 3D convolutional network aggregates. This aggregated correlation map is concatenated with the mean-intensity projection over the whole video and passed to a U-Net that predicts, for each pixel, a foreground/background score and affinities to nearby pixels. Affinities are converted to signed edge weights ($w_{ij}=a_{ij}-0.5$) and background edges are removed before the GASP algorithm partitions the graph into cell instances. On the Neurofinder test set, the paper reports that the per-series-trained variant (DISCos) achieves an average F1 of 0.67, outperforming all other methods trained only on the Neurofinder datasets, with 3dCNN at 0.66; the single-model variant (DISCo) reaches 0.63.

Load-bearing premise

The central outperformance claim depends on the Neurofinder leaderboard scores for other methods being current, accurate, and directly comparable; with a winning margin of only 0.01 average F1 over the runner-up 3dCNN, stale leaderboard entries or slight differences in training data would erase the claimed advantage.

Editorial extensions

If this is right

  • A single DISCo model, trained once on all five Neurofinder dataset series, attains a 0.63 average F1 on the test set, showing that one network can handle heterogeneous recordings without per-series tuning.
  • Per-series training (DISCos) raises the average F1 to 0.67 and edges out the previous best Neurofinder-trained method (3dCNN at 0.66), supporting the claim that the correlation-plus-shape design is competitive with spatiotemporal deep learning.
  • The single-video training experiment, with a 0.58 average F1 on test videos from the same series, implies that a neuroscience lab could label one video for a new recording setup and get usable segmentation without retraining on large data.
  • The lesion study's ordering of inputs implies that future systems should treat temporal correlations and shape summaries as complements rather than alternatives; dropping either cue costs roughly 0.13-0.14 in average F1.

Reading between the lines

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

  • None of the comparisons reports variance across training runs; given the 0.01 margin over 3dCNN, an obvious extension is to retrain DISCos with multiple seeds and report the distribution, which would tell whether the outperformance is systematic or within noise.
  • The affinity-plus-graph-partitioning recipe is agnostic to the imaging modality; the same signed-graph clustering of learned affinities has been used for other instance-segmentation tasks, so feeding temporal correlation features into that pipeline could transfer to video instance segmentation of other dynamic objects.
  • The authors do not analyze per-dataset precision/recall trade-offs; a testable extension would be to calibrate the background-edge exclusion threshold per dataset series, since the paper reports that recall limits series 00 and precision limits series 01.
  • Because the summary image carries the inactive cells, DISCo is positioned for all-cell segmentation; a natural next experiment is to evaluate it on datasets labeled specifically for active cells to see whether the correlation pathway can be dropped without loss.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 7 minor

Summary. The paper presents DISCo, a calcium-imaging cell segmentation method that computes segment-wise Pearson correlations between pixels, feeds these together with a mean intensity projection into a two-stage CNN that predicts pairwise affinities and a foreground/background map, and extracts instances by partitioning the resulting signed graph with the GASP algorithm. The method is evaluated on the Neurofinder public benchmark and compared with leaderboard entries, with additional lesion studies on input modalities and temporal aggregation. The main reported results are an average F1 of 0.63 for the single-model configuration DISCo and 0.67 for the per-series configuration DISCos, with the paper claiming that DISCos outperforms all other methods trained on the Neurofinder datasets.

Significance. If supported, the contribution would be valuable: DISCo offers a computationally cheaper alternative to 3D-CNN-on-video approaches while retaining temporal information, and it integrates instance segmentation directly. The code is publicly available, the evaluation is on an external public benchmark, and the lesion studies in Tables 4 and 5 support the main design choices. However, the central 'outperforms all' claim is not established as stated: it is misattributed in the abstract, it rests on an incomplete leaderboard comparison, and the 0.01 margin over 3dCNN is within the noise of single-run evaluations. With those claims corrected or tempered, the paper's contribution is solid and reproducible.

major comments (3)
  1. [Abstract and Section 1] The abstract states that 'DISCo outperforms all existing models trained on these datasets,' but Table 3 shows the single-model DISCo at average F1 0.63, below 3dCNN's 0.66; the 0.67 average that tops the listed Neurofinder-trained methods belongs to DISCos, the per-series configuration. Section 1 itself states the claim correctly for DISCos. The abstract and any unqualified excerpts should be revised so the top result is attributed to DISCos, and the Table 3 caption's statement about a 'single model' should be limited to comparisons among single-model methods.
  2. [Section 2 (footnote) and Table 3] The claim that DISCos 'outperforms all other methods trained on the Neurofinder datasets' is a universal claim over the leaderboard, but the comparison set is not demonstrated to be complete. The footnote in Section 2 excludes the Mask R-CNN and human-label submissions because no information is available on their models and training procedures; if either of these entries was trained only on Neurofinder training data and scored at least 0.67, the stated claim would be false. Since the leaderboard snapshot was taken on 2019-11-15, the claim should be restricted to the listed methods with verified training procedures and to that snapshot, or the training status of the excluded entries should be documented.
  3. [Section 4, Table 3] The reported advantage over the closest competitor 3dCNN is 0.01 in average F1 (0.67 vs. 0.66), based on single runs with no variance estimate and no significance test. A paired Wilcoxon signed-rank test on the nine per-video F1 scores in Table 3 yields p > 0.05, so the word 'outperforms' is not statistically supported for this comparison. The authors should either provide variance across multiple training runs with a significance test, or replace 'outperforms' by 'is competitive with' for comparisons involving 3dCNN. This is load-bearing because the headline claim rests on this margin.
minor comments (7)
  1. [Section 1] The word 'segmentaiton' in the second paragraph should be 'segmentation'.
  2. [Supplementary Table 7] In the DISCos/max row, '0,78' uses a comma as decimal separator; use '0.78' for consistency.
  3. [Figure 2 and Section 3.1] The figure shows C=3 neighbors while the text uses C=15; add a note that the figure is illustrative.
  4. [Section 3.3] The criterion for background exclusion, that the background prediction is higher than the foreground prediction, is equivalent to a 0.5 threshold on the foreground channel; state this explicitly.
  5. [Section 4] The exact p-values for the two-sided Wilcoxon tests against UNet2DS and Conv2D are not reported; give them in the text or in the table.
  6. [Table 3] Because the Neurofinder leaderboard is time-varying, the access date should be repeated in the caption of Table 3 rather than only in the body text.
  7. [Supplementary Table 8] Indicate that the missing standard deviations for some rows reflect series with only one training video, or explain why they are absent.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the benchmark evaluation is external, and DISCo's predictions do not reduce to its training inputs by construction.

full rationale

The paper's central claim is an empirical ranking on the Neurofinder benchmark. The network is trained on ground-truth affinities derived from the Neurofinder training annotations (Eq. 4) and evaluated on the undisclosed test set; no fitted parameter is renamed as a prediction, and the test F1 scores in Table 3 come from an external leaderboard. The only same-group citations are GASP [39], mutex watershed [38], and LeMoNADe [4], but GASP is used as a black-box graph-partitioning component whose linkage rule (Eq. 6) and stopping criterion are not fitted to Neurofinder test labels; the paper does not invoke a uniqueness theorem or import its conclusion from those citations. The exclusion of Mask R-CNN and human-label submissions in the Section 2 footnote, and the 0.01 average-F1 margin over 3dCNN, are concerns about claim completeness or statistical significance, not about circularity, and are therefore outside the scope of this pass. The method's components (segment-wise correlations, U-Net, GASP) are independently evaluated through the benchmark, so no derivation step is equivalent to its input by construction.

Assumptions & free parameters 8 free parameters · 4 assumptions · 0 invented entities

The main scientific content is an empirical ML pipeline; the central claim rests on benchmark labels, hyperparameters chosen on a validation split, and the implicit assumption that training and test videos have matched distributions. No new physical or mathematical entities are introduced.

free parameters (8)
  • N (number of temporal segments) = 10
    The video is split into 10 segments for segment-wise correlations; a hyperparameter chosen empirically.
  • C (number of correlation neighbors per pixel) = 15
    Correlations are computed to 15 neighboring pixels within distance 3; the neighborhood size was chosen empirically.
  • Graph edge connectivity distance = 5 pixels
    The signed graph connects pixels up to a distance of 5 to separate small adjacent cells; a design choice.
  • Minimum instance size threshold = 25 pixels
    Instances smaller than 25 pixels are removed as background in post-processing.
  • Inference max-pooling kernel length = 5 frames
    Temporal max-pooling is applied before correlations; the kernel length is fixed to 5 at inference and varied 3-9 during training.
  • Learning rate = 0.0001
    Adam optimizer learning rate selected via validation loss on a 25% spatial split of training videos.
  • Training epochs = 3000
    Number of epochs for DISCo and DISCos, selected via validation loss.
  • Batch size = 20 (DISCo); 6 for series 00 and 1 otherwise (DISCos)
    Batch sizes used for training; smaller batches for per-series networks due to less data.
assumptions (4)
  • domain assumption The Neurofinder ground truth annotations define the correct cell segmentation targets.
    Both training and evaluation rely on these labels; the paper notes labeling criteria differ across the five series (footnote 7), so noisy or inconsistent labels would skew reported F1 scores.
  • domain assumption Segment-wise Pearson correlations plus the mean intensity projection contain enough information to identify active and inactive cells.
    This is the core input design; it assumes cell pixel signals are correlated within cells after temporal splitting, and that silent cells are visible in the mean projection. The lesion study supports it empirically, but it is a domain assumption about the imaging modality.
  • domain assumption The 19 Neurofinder training videos are representative of the 9 test videos within each dataset series.
    The models are trained on the training set and evaluated on the test set; the supplementary single-video experiment reveals a systematic gap between training and test performance, indicating some distribution shift. The comparative claim still holds if all methods face the same shift, but absolute scores depend on this assumption.
  • domain assumption The GASP algorithm with average linkage is an appropriate solver for the affinity-based instance segmentation problem.
    The final instances come from GASP, a greedy signed-graph partitioning heuristic cited from the authors' own prior work [39]; the paper does not independently validate this choice beyond the end-to-end benchmark results.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DISCo: Deep learning, Instance Segmentation, and Correlations for cell segmentation in calcium imaging." pith.science (2026). https://pith.science/paper/ANFE5W6T

@misc{pith2026190807957,
  author       = {Pith},
  title        = {Pith review of: DISCo: Deep learning, Instance Segmentation, and Correlations for cell segmentation in calcium imaging},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ANFE5W6T}},
  note         = {Machine review of arXiv:1908.07957}
}
read the original abstract

Calcium imaging is one of the most important tools in neurophysiology as it enables the observation of neuronal activity for hundreds of cells in parallel and at single-cell resolution. In order to use the data gained with calcium imaging, it is necessary to extract individual cells and their activity from the recordings. We present DISCo, a novel approach for the cell segmentation in calcium imaging videos. We use temporal information from the recordings in a computationally efficient way by computing correlations between pixels and combine it with shape-based information to identify active as well as non-active cells. We first learn to predict whether two pixels belong to the same cell; this information is summarized in an undirected, edge-weighted grid graph which we then partition. In so doing, we approximately solve the NP-hard correlation clustering problem with a recently proposed greedy algorithm. Evaluating our method on the Neurofinder public benchmark shows that DISCo outperforms all existing models trained on these datasets.

Figures

Figures reproduced from arXiv: 1908.07957 by the authors.

Figure 1
Figure 1. DISCo workflow. For five exemplary pixel signals from the Neurofinder training set we show how they are processed [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. DISCo inputs and outputs. We show the inputs [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Example for the segment-wise correlations. We show the signals of two pixels [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Used edges for GASP. We show a small grid of [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 42 canonical work pages

  1. [1]

    Strickler, and Watt W

    Winfried Denk, James H. Strickler, and Watt W. Webb. Two- photon laser scanning fluorescence microscopy. Science, 1990

    work page 1990

  2. [2]

    Deep tissue two-photon microscopy

    Fritjof Helmchen and Winfried Denk. Deep tissue two-photon microscopy. Nature Methods, 2005

    work page 2005

  3. [3]

    High-speed, miniaturized fluorescence microscopy in freely moving mice

    Benjamin A Flusberg, Axel Nimmerjahn, Eric D Cocker, Eran A Mukamel, Robert PJ Barretto, Tony H Ko, Laurie D Burns, Juergen C Jung, and Mark J Schnitzer. High-speed, miniaturized fluorescence microscopy in freely moving mice. Nature Methods, 5(11):935, 2008

    work page 2008

  4. [4]

    Hamprecht

    Elke Kirschbaum, Manuel Haußmann, Steffen Wolf, Han- nah Sonntag, Justus Schneider, Shehabeldin Elzoheiry, Oliver Kann, Daniel Durstewitz, and Fred A. Hamprecht. LeMoN- ADe: Learned Motif and Neuronal Assembly Detection in calcium imaging videos. In International Conference on Learning Representations (ICLR), Proceedings, 2018

    work page 2018

  5. [5]

    Neurofinder public benchmark

    CodeNeuro. Neurofinder public benchmark. http: //neurofinder.codeneuro.org, 2016. Accessed: 2019-11-13

    work page 2016

  6. [6]

    Chet- tih, Matthias Minderer, Christopher Harvey, and Dorit S

    Quico Spaen, Roberto Asín-Achá, Selmaan N. Chet- tih, Matthias Minderer, Christopher Harvey, and Dorit S. Hochbaum. HNCcorr: A novel combinatorial ap- proach for cell identification in calcium-imaging movies. eNeuro, 6(2), 2019. doi: 10.1523/ENEURO.0304-18

    work page doi:10.1523/eneuro.0304-18 2019

  7. [7]

    Automated analysis of cellular signals from large-scale cal- cium imaging data

    Eran A Mukamel, Axel Nimmerjahn, and Mark J Schnitzer. Automated analysis of cellular signals from large-scale cal- cium imaging data. Neuron, 63(6):747–760, 2009

    work page 2009

  8. [8]

    Sparse non- negative deconvolution for compressive calcium imaging: algorithms and phase transitions

    Eftychios A Pnevmatikakis and Liam Paninski. Sparse non- negative deconvolution for compressive calcium imaging: algorithms and phase transitions. In Advances in Neural Information Processing Systems (NIPS), Proceedings, 2013

    work page 2013

Show all 43 references
  1. [9]

    Rank-penalized nonnegative spatiotemporal deconvolu- tion and demixing of calcium imaging data

    Eftychios A Pnevmatikakis, Timothy A Machado, Logan Grosenick, Ben Poole, Joshua T V ogelstein, and Liam Panin- ski. Rank-penalized nonnegative spatiotemporal deconvolu- tion and demixing of calcium imaging data. InComputational and Systems Neuroscience (Cosyne), 2013

    2013

  2. [10]

    Pnevmatikakis, Yuanjun Gao, Daniel Soudry, David Pfau, Clay Lacefield, Kira Poskanzer, Randy Bruno, Rafael Yuste, and Liam Paninski

    Eftychios A. Pnevmatikakis, Yuanjun Gao, Daniel Soudry, David Pfau, Clay Lacefield, Kira Poskanzer, Randy Bruno, Rafael Yuste, and Liam Paninski. A structured matrix fac- torization framework for large scale calcium imaging data analysis. arXiv preprint, arXiv:1409.2903, 2014

    2014 arXiv

  3. [11]

    Sparse space-time deconvolution for calcium image analysis

    Ferran Diego and Fred A Hamprecht. Sparse space-time deconvolution for calcium image analysis. In Advances in Neural Information Processing Systems (NIPS), Proceedings, 2014

    2014

  4. [12]

    Detecting cells using non-negative matrix factorization on calcium imaging data

    Ryuichi Maruyama, Kazuma Maeda, Hajime Moroda, Ichiro Kato, Masashi Inoue, Hiroyoshi Miyakawa, and Toru Aonishi. Detecting cells using non-negative matrix factorization on calcium imaging data. Neural Networks, 55:11–19, 2014

    2014

  5. [13]

    Pnevmatikakis, Daniel Soudry, Yuanjun Gao, Timothy A

    Eftychios A. Pnevmatikakis, Daniel Soudry, Yuanjun Gao, Timothy A. Machado, Josh Merel, David Pfau, Thomas Rear- don, Yu Mu, Clay Lacefield, Weijian Yang, Misha Ahrens, Randy Bruno, Thomas M. Jessell, Darcy S. Peterka, Rafael Yuste, and Liam Paninski. Simultaneous denoising, de...

    2016

  6. [14]

    Fast online deconvolution of calcium imaging data

    Johannes Friedrich, Pengcheng Zhou, and Liam Paninski. Fast online deconvolution of calcium imaging data. PLoS computational biology, 13(3):e1005423, 2017

    2017

  7. [15]

    Erdogdu, and Mark Schnitzer

    Hakan Inan, Murat A. Erdogdu, and Mark Schnitzer. Robust estimation of neural signals in calcium imaging. In Advances in Neural Information Processing Systems (NIPS), Proceed- ings, 2017

    2017

  8. [16]

    Onacid: Online analysis of cal- cium imaging data in real time

    Andrea Giovannucci, Johannes Friedrich, Matt Kaufman, Anne Churchland, Dmitri Chklovskii, Liam Paninski, and Eftychios A Pnevmatikakis. Onacid: Online analysis of cal- cium imaging data in real time. In Advances in Neural Infor- mation Processing Systems (NIPS), Proceedings, 2017

    2017

  9. [17]

    Efficient and accurate extraction of in vivo calcium signals from microen- doscopic video data

    Pengcheng Zhou, Shanna L Resendez, Jose Rodriguez- Romaguera, Jessica C Jimenez, Shay Q Neufeld, Andrea Giovannucci, Johannes Friedrich, Eftychios A Pnevmatikakis, Garret D Stuber, Rene Hen, Mazen A Kheirbek, Bernardo L Sabatini, Robert E Kass, and Liam Paninski. Efficient and ...

    2018

  10. [18]

    Caiman an open source tool for scalable calcium imaging data analysis

    Andrea Giovannucci, Johannes Friedrich, Pat Gunn, Jérémie Kalfon, Brandon L Brown, Sue Ann Koay, Jiannis Taxidis, Farzaneh Najafi, Jeffrey L Gauthier, Pengcheng Zhou, Baljit S Khakh, David W Tank, Dmitri B Chklovskii, and Eftychios A Pnevmatikakis. Caiman an open source tool fo...

    2019

  11. [19]

    SIMA: Python software for analysis of dynamic fluorescence imaging data

    Patrick Kaifosh, Jeffrey D Zaremba, Nathan B Danielson, and Attila Losonczy. SIMA: Python software for analysis of dynamic fluorescence imaging data. Frontiers in neuroinfor- matics, 8:80, 2014

    2014

  12. [20]

    Hamprecht

    Ferran Diego, Susanne Reichinnek, Martin Both, and Fred A. Hamprecht. Automated identification of neuronal activity from calcium imaging by sparse dictionary learning. In IEEE International Symposium on Biomedical Imaging (ISBI), 2013

    2013

  13. [21]

    Learning multi-level sparse representations

    Ferran Diego and Fred A Hamprecht. Learning multi-level sparse representations. In Advances in Neural Information Processing Systems (NIPS), Proceedings, 2013

    2013

  14. [22]

    Extracting regions of interest from biological images with convolutional sparse block coding

    Marius Pachitariu, Adam M Packer, Noah Pettit, Henry Dal- gleish, Michael Hausser, and Maneesh Sahani. Extracting regions of interest from biological images with convolutional sparse block coding. In Advances in Neural Information Processing Systems (NIPS), Proceedings, 2013

    2013

  15. [23]

    Scalpel: Extracting neurons from calcium imaging data

    Ashley Petersen, Noah Simon, and Daniela Witten. Scalpel: Extracting neurons from calcium imaging data. The annals of applied statistics, 12(4):2430, 2018

    2018

  16. [24]

    Apthorpe, Alexander J

    Noah J. Apthorpe, Alexander J. Riordan, Rob E. Aguilar, Jan Homann, Yi Gu, David W. Tank, and H. Sebastian Seung. Automatic neuron detection in calcium imaging data using convolutional networks. In Advances in Neural Information Processing Systems (NIPS), Proceedings, 2016

    2016

  17. [25]

    Fast, simple calcium imaging segmentation with fully convolutional networks

    Aleksander Klibisz, Derek Rose, Matthew Eicholtz, Jay Blun- don, and Stanislav Zakharenko. Fast, simple calcium imaging segmentation with fully convolutional networks. In Interna- tional Workshop on Deep Learning in Medical Image Analy- sis and Multimodal Learning for Clinical...

    2017

  18. [26]

    Fast and robust active neu- ron segmentation in two-photon calcium imaging using spa- tiotemporal deep learning

    Somayyeh Soltanian-Zadeh, Kaan Sahingur, Sarah Blau, Yiyang Gong, and Sina Farsiu. Fast and robust active neu- ron segmentation in two-photon calcium imaging using spa- tiotemporal deep learning. Proceedings of the National Academy of Sciences , 116(17):8554–8563, 2019. ISSN 0...

    2019 doi

  19. [27]

    Shang Gao. Conv2d. https://github.com/ iamshang1/Projects/tree/master/Advanced_ ML/Neuron_Detection, 2016. Accessed: 2019-07-19

    2016

  20. [28]

    Suite2p: beyond 10,000 neu- rons with standard two-photon microscopy

    Marius Pachitariu, Carsen Stringer, Mario Dipoppa, Sylvia Schröder, L Federico Rossi, Henry Dalgleish, Matteo Caran- dini, and Kenneth D Harris. Suite2p: beyond 10,000 neu- rons with standard two-photon microscopy. bioRxiv preprint, bioRxiv:061507, 2017

    2017

  21. [29]

    Allen Brain Observatory (ABO) datasets

    Allen Institute for Brain Science. Allen Brain Observatory (ABO) datasets. http://observatory.brain-map. org/visualcoding, 2016. Accessed: 2019-08-20

    2016

  22. [30]

    Scaled correlation analysis: a better way to compute a cross-correlogram

    Danko Nikoli ´c, Raul C Mure¸ san, Weijia Feng, and Wolf Singer. Scaled correlation analysis: a better way to compute a cross-correlogram. European Journal of Neuroscience, 35 (5):742–762, 2012

    2012

  23. [31]

    Synchronisation hubs in the visual cor- tex may arise from strong rhythmic inhibition during gamma oscillations

    Stefanos E Folias, Shan Yu, Abigail Snyder, Danko Nikoli´c, and Jonathan E Rubin. Synchronisation hubs in the visual cor- tex may arise from strong rhythmic inhibition during gamma oscillations. European Journal of Neuroscience, 38(6):2864– 2883, 2013

    2013

  24. [32]

    A scaled- correlation based approach for defining and analyzing func- tional networks

    Samuel Dolean, Mihaela Dîn¸ soreanu, Raul Cristian Mure¸ san, Attila Geiszt, Rodica Potolea, and Ioana ¸ Tinca¸ s. A scaled- correlation based approach for defining and analyzing func- tional networks. In International Workshop on New Frontiers in Mining Complex Patterns, pages...

    2017

  25. [33]

    Notes on regression and inheritance in the case of two parents

    Karl Pearson. Notes on regression and inheritance in the case of two parents. Proceedings of the Royal Society of London, 58:240–242, 1895

  26. [34]

    Measur- ing and testing dependence by correlation of distances

    Gábor J Székely, Maria L Rizzo, and Nail K Bakirov. Measur- ing and testing dependence by correlation of distances. The annals of statistics, 35(6):2769–2794, 2007

    2007

  27. [35]

    U-net: Convolutional networks for biomedical image segmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), pages 234– 241, 2015

    2015

  28. [36]

    Measures of the amount of ecologic association between species

    Lee R Dice. Measures of the amount of ecologic association between species. Ecology, 26(3):297–302, 1945

    1945

  29. [37]

    A method of establishing groups of equal amplitude in plant sociology based on similarity of species content and its application to analyses of the vegeta- tion on danish commons

    Thorvald Julius Sørensen. A method of establishing groups of equal amplitude in plant sociology based on similarity of species content and its application to analyses of the vegeta- tion on danish commons. Biol. Skar., 5:1–34, 1948

    1948

  30. [38]

    Ham- precht

    Steffen Wolf, Constantin Pape, Alberto Bailoni, Nasim Ra- haman, Anna Kreshuk, Ullrich Köthe, and Fred A. Ham- precht. The mutex watershed: efficient, parameter-free image partitioning. In European Conference on Computer Vision (ECCV), Proceedings, 2018

    2018

  31. [39]

    Hamprecht

    Alberto Bailoni, Constantin Pape, Steffen Wolf, Thorsten Beier, Anna Kreshuk, and Fred A. Hamprecht. A general- ized framework for agglomerative clustering of signed graphs applied to instance segmentation. arXiv preprint, arXiv: 1906.11713, 2019

    1906 arXiv

  32. [40]

    Efficient decomposition of image and mesh graphs by lifted multi- cuts

    Margret Keuper, Evgeny Levinkov, Nicolas Bonneel, Guil- laume Lavoué, Thomas Brox, and Bjorn Andres. Efficient decomposition of image and mesh graphs by lifted multi- cuts. In IEEE International Conference on Computer Vision (ICCV), Proceedings, 2015

    2015

  33. [41]

    A comparative study of local search algorithms for correlation clustering

    Evgeny Levinkov, Alexander Kirillov, and Bjoern Andres. A comparative study of local search algorithms for correlation clustering. In German Conference on Pattern Recognition, pages 103–114. Springer, 2017

    2017

  34. [42]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In International Conference on Learning Representations (ICLR), Proceedings, 2014. Supplementary Material Table 6: Results with DISCo for different kinds of input. We show the results on the Neurofinder...

    2014

  35. [2019]

    URL https://www.eneuro.org/content/ 6/2/ENEURO.0304-18.2019

    2019

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.