analiza aiqbench
This research introduces DPNeXt, a lightweight decoder framework designed to optimize multi-task learning for robotics perception by efficiently leveraging frozen Vision Foundation Models. By employing fusion-centric decoding and novel boundary guidance strategies, the method significantly reduces parameter count while maintaining state-of-the-art performance in semantic segmentation and depth estimation on resource-constrained hardware. The work addresses critical bottlenecks in current dense prediction pipelines without relying on extra annotations or inference overhead.
DPNeXt: A Lightweight Multi-Scale Feature Fusion Framework for Efficient ViT-Based Multi-Task Dense Prediction
arXiv:2607.16012v1 Announce Type: cross Abstract: Multi-Task Learning (MTL) in robotics perception systems supports comprehensive 3D spatial scene understanding by integrating semantic segmentation and depth estimation. While Vision Foundation Models (VFMs) are increasingly adopted as robust feature encoders, existing decoding strategies present a critical bottleneck. To address this, we propose DPNeXt, a streamlined multi-scale feature fusion decoder and efficient alternative to the standard Dense Prediction Transformer (DPT). DPNeXt uses dual depthwise separable inverted bottlenecks to improve frozen VFM utilization through fusion-centric decoding and independent task modularization. To further mitigate negative inductive transfer between tasks, we introduce the Multi-Task Boundary Guidance (MTBG) strategy. Unlike prior boundary-aware methods that add fusion modules or gating, MTBG applies symmetric boundary-focused supervision to encourage geometric consistency without extra annotation or inference cost. Experiments on Cityscapes show that DPNeXt-S outperforms prior state-of-the-art (SOTA) MTL models, while DPNeXt-B further improves the overall performance and achieves the best results among the compared methods. On NYUv2, DPNeXt-B also achieves the best semantic segmentation and depth estimation results among the compared methods while requiring substantially fewer trainable parameters than prior large-scale MTL models. Compared with the standard DPT, DPNeXt-S reduces trainable parameters by 78.6% and achieves the fastest inference speed among the compared models on resource-constrained laptop hardware. The source code, model checkpoints, and a demo video will be made available at https://github.com/kangjehun/DPNeXt.
Abstract:Multi-Task Learning (MTL) in robotics perception systems supports comprehensive 3D spatial scene understanding by integrating semantic segmentation and depth estimation. While Vision Foundation Models (VFMs) are increasingly adopted as robust feature encoders, existing decoding strategies present a critical bottleneck. To address this, we propose DPNeXt, a streamlined multi-scale feature fusion decoder and efficient alternative to the standard Dense Prediction Transformer (DPT). DPNeXt uses dual depthwise separable inverted bottlenecks to improve frozen VFM utilization through fusion-centric decoding and independent task modularization. To further mitigate negative inductive transfer between tasks, we introduce the Multi-Task Boundary Guidance (MTBG) strategy. Unlike prior boundary-aware methods that add fusion modules or gating, MTBG applies symmetric boundary-focused supervision to encourage geometric consistency without extra annotation or inference cost. Experiments on Cityscapes show that DPNeXt-S outperforms prior state-of-the-art (SOTA) MTL models, while DPNeXt-B further improves the overall performance and achieves the best results among the compared methods. On NYUv2, DPNeXt-B also achieves the best semantic segmentation and depth estimation results among the compared methods while requiring substantially fewer trainable parameters than prior large-scale MTL models. Compared with the standard DPT, DPNeXt-S reduces trainable parameters by 78.6% and achieves the fastest inference speed among the compared models on resource-constrained laptop hardware. The source code, model checkpoints, and a demo video will be made available at this https URL.
8 pages, 5 figures. Accepted to the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2026)
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Robotics (cs.RO)
arXiv:2607.16012 [cs.CV]
(or arXiv:2607.16012v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2607.16012 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.
Text integral sub licența deschisă a publicației · 339 cuvinte
Subiecte conexe
News
Quantum Computing Report · 25 iul. 2026
SGP-SAM: Self-Gated Prompting for Transferring 3D Segment Anything Models to Lesion Segmentation
arXiv · 29 sept. 2026
IndoorBEV: A Lightweight Real-Time LiDAR BEV Perception System for Indoor Mobile Robots
arXiv · 2 oct. 2026
ActionUNet: Improving Robustness of VLA Models with Efficient Multi-scale Fine-tuning
arXiv · 29 sept. 2026