Source author record

Kan Yang

Kan Yang appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Computer Vision Information Theory math.IT Networking and Internet Architecture

Catalog footprint

What is connected

2works

4topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Last-Layer-Centric Feature Recombination: Unleashing 3D Geometric Knowledge in DINOv3 for Monocular Depth Estimation

Monocular depth estimation (MDE) is a fundamental yet inherently ill-posed task. Recent vision foundation models (VFMs), particularly DINO-based transformers, have significantly improved accuracy and generalization for dense prediction. Prior works generally follow a unified paradigm: sampling a fixed set of intermediate transformer layers at uniform intervals to build multi-scale features. This common practice implicitly assumes that geometric information is uniformly distributed across layers, which may underutilize the structural 3D cues encoded in VFMs. In this study, we present a systematic layer-wise analysis of DINOv3, revealing that 3D information is distributed non-uniformly: deeper layers exhibit stronger depth predictability and better capture inter-sample geometric variation. Motivated by this, we introduce a Last-Layer-Centric Feature Recombination (LFR) module to enhance geometric expressiveness. LFR treats the final layer as a geometric anchor and adaptively selects complementary intermediate layers according to a minimal-similarity criterion. Selected features are fused with the last-layer representation via compact linear adapters.Extensive experiments show that LFR module consistently improves MDE accuracy and achieves state-of-the-art performance. Our analysis sheds light on how geometric knowledge is organized within VFMs and offers an efficient strategy for unlocking their potential in dense 3D tasks.

preprint2016arXiv

Optimal Reliability in Energy Harvesting Industrial Wireless Sensor Networks

For Industrial Wireless Sensor Networks, it is essential to reliably sense and deliver the environmental data on time to avoid system malfunction. While energy harvesting is a promising technique to extend the lifetime of sensor nodes, it also brings new challenges for system reliability due to the stochastic nature of the harvested energy. In this paper, we investigate the optimal energy management policy to minimize the weighted packet loss rate under delay constraint, where the packet loss rate considers the lost packets both during the sensing and delivering processes. We show that the above energy management problem can be modeled as an infinite horizon average reward constraint Markov decision problem. In order to address the well-known curse of dimensionality problem and facilitate distributed implementation, we utilize the linear value approximation technique. Moreover, we apply stochastic online learning with post-decision state to deal with the lack of knowledge of the underlying stochastic processes. A distributed energy allocation algorithm with water-filling structure and a scheduling algorithm by auction mechanism are obtained. Experimental results show that the proposed algorithm achieves nearly the same performance as the optimal offline value iteration algorithm while requiring much less computation complexity and signaling overhead, and outperforms various existing baseline algorithms.