From the Drosophila Visual Connectome to General-Purpose Computer Vision

Abstract

Biological connectomes encode structured solutions to visual computation that may provide reusable inductive biases for artificial vision. We develop ConnectomeX around FlyVision, a trainable architecture that preserves parallel ON/OFF processing, recurrent computation and population-level graph interaction while scaling model capacity across tasks. FlyVision reached 99.34% accuracy on MNIST with 80,608 parameters and 78.03% on CIFAR-10 with 81,408 parameters. On ImageNet-1K, FlyVision Base and Large reached 60.79% and 66.25% top-1 accuracy with 1.8 and 3.7 million parameters, while a Large local-k7 model with a learned low-frequency branch reached 66.53%, compared with 69.25% for ResNet18 with 11.7 million parameters. On a 22-class skin-disease benchmark, FlyVision Large achieved 63.78% accuracy and 95.28% macro-AUROC with 2.99 million parameters. In four-class chest radiography, ImageNet-pretrained FlyVision Base and Large reached 92.60% and 92.76% accuracy with 1.33 and 2.97 million parameters, compared with 91.56% for ImageNet-pretrained ResNet18 with 11.18 million. BrainAGE extends FlyVision to volumetric T1-weighted MRI by applying a shared ImageNet-pretrained FlyVision Large encoder to 24 sagittal, coronal and axial slices per scan and combining slice-level age estimates by confidence-modulated Gaussian voting. On 433 held-out scans, three-axis fusion achieved a mean absolute error of 5.98 years and R^2 = 0.868. Across the 224x224 classification tasks, the best FlyVision configuration remained within three percentage points of ResNet18 on ImageNet-1K and skin-disease classification and exceeded it on chest radiography with substantially fewer parameters. These results show that a conserved connectome-informed computation can scale from compact recognition to large-scale natural and biomedical vision.

Connectome analysis reveals brain-wide visual processing in Drosophila
Sung Yong Kim, Anmo J. Kim
iScience1 Nov 2026Peer-reviewedCC BY-NC 4.0NeuroscienceCrossrefCell Press
Representational geometry as a fidelity metric for connectome-constrained networks: evidence from the Drosophila visual system
M. G. Zhou, J. O. Hasler
bioRxiv10 Jun 2026PreprintCC BYFull textNeurosciencebioRxiv
The complete gustatory connectome of adult Drosophila reveals how taste guides feeding, foraging, and social behavior
Ibrahim Tastekin, Ines de Haan Vicente, Rory J. Beresford +10
Cell1 Sep 2026Peer-reviewedLicence not statedFull text6 citedNeuroscienceOpenAlex
A canonical excitatory-inhibitory motif embedded in the connectome underlies rhythmic locomotion in Drosophila larvae
T. Seki, H. Kohsaka, M. F. Zwart +1
bioRxiv31 Aug 2026PreprintCC BY-NC-NDFull textNeurosciencebioRxiv
GGE: General-purpose deep meta-learning for classification of human transcriptomes with limited data
G. Yankovitz, I. Gat-Viks
bioRxiv27 Sep 2026PreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv

Cite

BibTeX RIS

BibTeX

@misc{li2026from,
  title = {{From the Drosophila Visual Connectome to General-Purpose Computer Vision}},
  author = {Zongyu Li and Akito Yamauchi and Huaizhi Liu and Vishwanatha Rao and Jia Guo and for the Frontotemporal Lobar Degeneration Neuroimaging Initiative and for the Alzheimer's Disease Neuroimaging Initiative},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.08418},
  eprint = {2610.08418},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.08418},
}

RIS

TY  - UNPB
TI  - From the Drosophila Visual Connectome to General-Purpose Computer Vision
AU  - Zongyu Li
AU  - Akito Yamauchi
AU  - Huaizhi Liu
AU  - Vishwanatha Rao
AU  - Jia Guo
AU  - for the Frontotemporal Lobar Degeneration Neuroimaging Initiative
AU  - for the Alzheimer's Disease Neuroimaging Initiative
PY  - 2026
DA  - 2026/10/06
PB  - arXiv
DO  - 10.48550/arxiv.2610.08418
UR  - https://arxiv.org/abs/2610.08418
AB  - Biological connectomes encode structured solutions to visual computation that may provide reusable inductive biases for artificial vision. We develop ConnectomeX around FlyVision, a trainable architecture that preserves parallel ON/OFF processing, recurrent computation and population-level graph interaction while scaling model capacity across tasks. FlyVision reached 99.34% accuracy on MNIST with 80,608 parameters and 78.03% on CIFAR-10 with 81,408 parameters. On ImageNet-1K, FlyVision Base and Large reached 60.79% and 66.25% top-1 accuracy with 1.8 and 3.7 million parameters, while a Large local-k7 model with a learned low-frequency branch reached 66.53%, compared with 69.25% for ResNet18 with 11.7 million parameters. On a 22-class skin-disease benchmark, FlyVision Large achieved 63.78% accuracy and 95.28% macro-AUROC with 2.99 million parameters. In four-class chest radiography, ImageNet-pretrained FlyVision Base and Large reached 92.60% and 92.76% accuracy with 1.33 and 2.97 million parameters, compared with 91.56% for ImageNet-pretrained ResNet18 with 11.18 million. BrainAGE extends FlyVision to volumetric T1-weighted MRI by applying a shared ImageNet-pretrained FlyVision Large encoder to 24 sagittal, coronal and axial slices per scan and combining slice-level age estimates by confidence-modulated Gaussian voting. On 433 held-out scans, three-axis fusion achieved a mean absolute error of 5.98 years and R^2 = 0.868. Across the 224x224 classification tasks, the best FlyVision configuration remained within three percentage points of ResNet18 on ImageNet-1K and skin-disease classification and exceeded it on chest radiography with substantially fewer parameters. These results show that a conserved connectome-informed computation can scale from compact recognition to large-scale natural and biomedical vision.
ER  -

Abstract

Related

Cite