Salvato in:
Dettagli Bibliografici
Autore principale: Stolle, Kurt H. W.
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2412.07966
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917865880289280
author Stolle, Kurt H. W.
author_facet Stolle, Kurt H. W.
contents In this work, we present Multiformer, a novel approach to depth-aware video panoptic segmentation (DVPS) based on the mask transformer paradigm. Our method learns object representations that are shared across segmentation, monocular depth estimation, and object tracking subtasks. In contrast to recent unified approaches that progressively refine a common object representation, we propose a hybrid method using task-specific branches within each decoder block, ultimately fusing them into a shared representation at the block interfaces. Extensive experiments on the Cityscapes-DVPS and SemKITTI-DVPS datasets demonstrate that Multiformer achieves state-of-the-art performance across all DVPS metrics, outperforming previous methods by substantial margins. With a ResNet-50 backbone, Multiformer surpasses the previous best result by 3.0 DVPQ points while also improving depth estimation accuracy. Using a Swin-B backbone, Multiformer further improves performance by 4.0 DVPQ points. Multiformer also provides valuable insights into the design of multi-task decoder architectures.
format Preprint
id arxiv_https___arxiv_org_abs_2412_07966
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Balancing Shared and Task-Specific Representations: A Hybrid Approach to Depth-Aware Video Panoptic Segmentation
Stolle, Kurt H. W.
Computer Vision and Pattern Recognition
In this work, we present Multiformer, a novel approach to depth-aware video panoptic segmentation (DVPS) based on the mask transformer paradigm. Our method learns object representations that are shared across segmentation, monocular depth estimation, and object tracking subtasks. In contrast to recent unified approaches that progressively refine a common object representation, we propose a hybrid method using task-specific branches within each decoder block, ultimately fusing them into a shared representation at the block interfaces. Extensive experiments on the Cityscapes-DVPS and SemKITTI-DVPS datasets demonstrate that Multiformer achieves state-of-the-art performance across all DVPS metrics, outperforming previous methods by substantial margins. With a ResNet-50 backbone, Multiformer surpasses the previous best result by 3.0 DVPQ points while also improving depth estimation accuracy. Using a Swin-B backbone, Multiformer further improves performance by 4.0 DVPQ points. Multiformer also provides valuable insights into the design of multi-task decoder architectures.
title Balancing Shared and Task-Specific Representations: A Hybrid Approach to Depth-Aware Video Panoptic Segmentation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.07966