ACCOR: Attention-Enhanced Complex-Valued Contrastive Learning for Occluded Object Classification Using mmWave Radar IQ Signals
arXiv:2512.11556v4 Announce Type: replace Abstract: Millimeter-wave (mmWave) radar provides robust sensing under adverse conditions and can penetrate thin materials for non-visual perception in industrial and robotic settings. Recent work with MIMO mmWave radar has demonstrated its ability to penetrate cardboard packaging for occluded object classification. However, existing models leave room for improvement and extensions across different sensing frequencies. Building on recent work with MIMO radar for occluded object classification, we propose ACCOR, an attention-enhanced complex-valued contrastive learning approach for radar, enabling robust occluded object classification. ACCOR processes complex-valued IQ radar signals via a complex-valued CNN backbone, a multi-head attention layer and
View PDF
Abstract:Millimeter-wave (mmWave) radar provides robust sensing under adverse conditions and can penetrate thin materials for non-visual perception in industrial and robotic settings. Recent work with MIMO mmWave radar has demonstrated its ability to penetrate cardboard packaging for occluded object classification. However, existing models leave room for improvement and extensions across different sensing frequencies. Building on recent work with MIMO radar for occluded object classification, we propose ACCOR, an attention-enhanced complex-valued contrastive learning approach for radar, enabling robust occluded object classification. ACCOR processes complex-valued IQ radar signals via a complex-valued CNN backbone, a multi-head attention layer and a hybrid loss. The hybrid loss combines a weighted cross-entropy term with a supervised contrastive term. We extend an existing 64 GHz dataset with a new 67 GHz subset and evaluate performance across both bands. ACCOR achieves 96.60 % accuracy at 64 GHz and 93.59 % at 67 GHz on 10 objects, surpassing prior radar-specific and adapted image models. Results demonstrate the benefits of integrating complex-valued deep learning, attention, and contrastive learning for mmWave radar-based occluded object classification.
Comments: 9 pages, 8 figures
Subjects:
Signal Processing (eess.SP)
Cite as: arXiv:2512.11556 [eess.SP]
(or arXiv:2512.11556v4 [eess.SP] for this version)
https://doi.org/10.48550/arXiv.2512.11556
arXiv-issued DOI via DataCite
Submission history
From: Stefan Hägele [view email] [v1] Fri, 12 Dec 2025 13:38:59 UTC (1,650 KB) [v2] Thu, 5 Mar 2026 13:28:30 UTC (1,980 KB) [v3] Fri, 6 Mar 2026 13:43:52 UTC (1,979 KB) [v4] Thu, 2 Apr 2026 08:19:11 UTC (2,195 KB)
Sign in to highlight and annotate this article
Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
modelannouncearxiv
60% of Consumers Want Approval Gates for AI Spending. Who Builds Them?
Visa just published a study of 2,000 consumers on AI agents and spending. The finding that should dominate every conversation about agentic commerce: 60% of respondents want human approval gates before an AI agent makes purchases on their behalf. Only 27% are comfortable with unlimited AI spending authority. Thirty-six percent say they would trust an AI agent backed by their bank. Twenty-eight percent would trust an independent agent. The paper's own summary: "Trust is the adoption switch." This is empirical confirmation of something that was structurally obvious. The infrastructure to move money is almost ready. The infrastructure to decide whether money should move does not exist. The asymmetry Two days ago, the x402 Foundation launched under the Linux Foundation. Twenty-two founding mem
Autonomous Revolution: How AI Agents are Redefining Blockchain's Future
Autonomous Revolution: How AI Agents are Redefining Blockchain's Future The convergence of artificial intelligence and blockchain technology is no longer a futuristic pipe dream; it's a rapidly unfolding reality. As developers and innovators push the boundaries of decentralized systems, the integration of AI agents in blockchain is emerging as a critical catalyst for enhanced efficiency, security, and user experience. This article delves into the transformative potential of these intelligent entities, exploring how they are poised to reshape the crypto landscape and unlock unprecedented levels of automation and intelligence within Web3. The Symbiotic Relationship: AI Enhancing Decentralization Blockchain's core tenets of decentralization, immutability, and transparency are powerful, but th
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!