aiqbench analysis
This research advances computer vision by demonstrating the efficacy of Vision Transformer ensembles for panoramic street segmentation, achieving top-tier performance in a competitive benchmark while maintaining strict computational budgets. The study provides critical insights into model selection and reproducibility workflows, comparing established architectures like DeepLabV3+ against specialized Mask Transformers to optimize inference efficiency on modern GPU hardware.
Vision Transformer Ensembles for Panoramic Street Segmentation
arXiv:2610.06063v1 Announce Type: cross Abstract: Semantic segmentation of street panoramas can support detailed descriptions of urban environments, yet small datasets and unequal training costs make model selection difficult. This paper presents the system used for a first place submission to the PalmCity challenge in the leaderboard snapshot dated 5 October 2026. Nine pretrained segmentation systems are compared using approximately equal computation budgets. The candidates include DeepLabV3+, SegFormer, UPerNet, Mask2Former, DINOv3 with a linear decoder, and an Encoder only Mask Transformer using DINOv3. The two leading candidates are trained independently with three random seeds and longer budgets. Equal averaging of class probabilities from the three Encoder only Mask Transformer models, evaluated at three image scales with horizontal reflection, produces 60.95% mean intersection over union and 71.16% mean F1 on the 84 image public validation split. The submitted predictions receive 57.08% mean intersection over union and 67.96% mean F1 on the hidden test leaderboard. Producing all 249 test masks takes 251.49 seconds including model initialization and provenance checks on one NVIDIA RTX 5090. Peak allocated GPU memory is 2.70 GiB. The study reports all eligible models, all inference variants, class level errors, source conditions, and reproducibility checks, providing a documented challenge workflow with existing architectures.
Abstract:Semantic segmentation of street panoramas can support detailed descriptions of urban environments, yet small datasets and unequal training costs make model selection difficult. This paper presents the system used for a first place submission to the PalmCity challenge in the leaderboard snapshot dated 5 October 2026. Nine pretrained segmentation systems are compared using approximately equal computation budgets. The candidates include DeepLabV3+, SegFormer, UPerNet, Mask2Former, DINOv3 with a linear decoder, and an Encoder only Mask Transformer using DINOv3. The two leading candidates are trained independently with three random seeds and longer budgets. Equal averaging of class probabilities from the three Encoder only Mask Transformer models, evaluated at three image scales with horizontal reflection, produces 60.95% mean intersection over union and 71.16% mean F1 on the 84 image public validation split. The submitted predictions receive 57.08% mean intersection over union and 67.96% mean F1 on the hidden test leaderboard. Producing all 249 test masks takes 251.49 seconds including model initialization and provenance checks on one NVIDIA RTX 5090. Peak allocated GPU memory is 2.70 GiB. The study reports all eligible models, all inference variants, class level errors, source conditions, and reproducibility checks, providing a documented challenge workflow with existing architectures.
18 pages, 4 figures. Code available at this https URL . Trained models available at this https URL
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2610.06063 [cs.CV]
(or arXiv:2610.06063v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2610.06063 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 319 words
Related coverage
Post-Operative Glioma Segmentation via Loss Stabilization, Normalization and Subspace Attention
arXiv · Jul 28, 2026
Enhancing Road Safety Through Multi-Camera Image Segmentation with Post-Encroachment Time Analysis
arXiv · Sep 3, 2026
We’re Squandering LEDs’ Potential to Save Our Night Skies
IEEE Spectrum · Jul 20, 2026
Colorectal Cancer Segmentation with Adaptive Augmentation and Multi-Resolution Ensemble Models
arXiv · Oct 1, 2026