CVPR logo GRAIL-V logo

CALIBRA: Calibration-Aware Multi-Agent Verification for Contactless Physiological Monitoring

1Department of Information Systems, University of Maryland, Baltimore County, USA
Accepted Long paper at CVPR 2026 (GRAIL-V)

34.4%

MAE reduction over zero-shot prompting with Gemini-2.5-Flash

31.0%

MAE reduction over zero-shot prompting with GPT-4o

93%

Coverage in routed mode at 4.1 s average latency and 1.6 GB peak memory

Figure 1

CALIBRA improves over one-shot prompting by routing difficult segments, checking agreement across evidence channels, repairing conflicts when needed, and abstaining when evidence remains unreliable.

Abstract

Contactless physiological monitoring offers a promising path toward low-burden health sensing, but reliable inference from short video segments remains difficult when visual evidence is weak, corrupted by motion, or spectrally ambiguous. We present CALIBRA, a calibration-aware multi-agent framework for contactless respiratory rate (RR) estimation that treats prediction as an evidence-grounded verification problem rather than a one-shot prompting task. CALIBRA operates on a structured artifact bundle with lightweight signal-derived candidate rates, and performs routing, specialized vision-language evidence analysis, critic-based consistency checking, targeted repair, and calibrated arbitration with optional abstention. On our in-house benchmark, CALIBRA consistently outperforms prompting-only baselines across multiple backbones. With Gemini-2.5-Flash, it reduces mean absolute error (MAE) from 5.21 to 3.42, corresponding to a 34.4% reduction over zero-shot prompting and a 16.8% reduction over chain-of-thought prompting. With GPT-4o, it reduces MAE from 5.48 to 3.78, a 31.0% reduction over zero-shot prompting. CALIBRA also improves grounding and consistency in judge-based evaluation, indicating better evidence alignment beyond numerical accuracy alone. Under resource constraints, a routed configuration maintains 93% coverage at 4.1 s average latency and 1.6 GB peak memory. These results show that robust contactless physiological inference depends not only on stronger multimodal backbones, but also on explicit verification of whether predictions are supported by the available evidence.

CALIBRA is a calibration-aware multi-agent framework for contactless respiratory rate estimation. Instead of relying on one-shot prompting alone, it combines routing, evidence verification, repair, and calibrated arbitration to produce more reliable predictions under uncertain visual evidence.

Why CALIBRA

Prompt-only inference can fail quietly

Difficult segments often contain motion corruption, weak periodicity, or harmonic ambiguity. A model can still output a plausible respiratory rate even when the evidence is inconsistent.

Verification improves decision discipline

CALIBRA checks candidate rates against structured evidence, selectively repairs conflicts, and commits only after consistency improves.

Reliability matters in deployment

The framework can abstain when evidence is not coherent enough, improving calibration and safety instead of forcing unreliable predictions.

Method

Each segment is represented as a four-view artifact bundle with an annotated frame, respiratory signal plot, optical-flow summary, and histogram view, together with deterministic candidate respiratory rates. The framework first estimates whether a segment appears easy or ambiguous, then routes it through specialized agents that inspect waveform structure, motion quality, visual periodicity, and disagreement between candidate estimates. A critic module aggregates the evidence, checks consistency, and flags conflicts. When needed, a repair stage revisits uncertain evidence and updates the preferred interpretation before an arbiter produces the final respiratory-rate estimate together with a calibrated confidence decision. This design allows CALIBRA to move beyond one-pass prompting and instead make decisions through explicit evidence checking and controlled refinement.

Figure 2

Overview of CALIBRA’s evidence-grounded verification pipeline.

Model's Performance

CALIBRA improves error, grounding, and consistency while remaining practical under routed deployment.
3.42
MAE
Gemini-2.5-Flash + CALIBRA
34.4%
Reduction
vs zero-shot prompting
31.0%
Reduction
GPT-4o vs zero-shot
93%
Coverage
routed mode

Main benchmark results

lower is better
Method MAE ↓ RMSE ↓ Bias ↓
CALIBRA (Gemini-2.5-Flash) 3.42 4.88 0.29
CALIBRA (GPT-4o) 3.78 5.26 0.34
CALIBRA (Qwen3-VL-8B) 4.35 5.92 0.41
CALIBRA (DeepSeek-R1-7B) 4.58 6.15 0.45
CALIBRA (InternVL3-8B) 4.71 6.28 0.47
0.62
RG
response grounding
0.79
CRA
critic agreement
0.90
CS
consistency score
0.82
EAR
evidence alignment

Judge-based output quality

higher is better
Model / Setting Grounding ↑ Consistency ↑ Rationale ↑ Confidence ↑ Plausibility ↑
GPT-4o (CoT) 7.6 7.4 7.3 7.0 7.5
GPT-4o (CALIBRA) 8.6 8.4 8.2 8.1 8.5
Qwen3-VL-8B (CoT) 7.1 6.9 6.8 6.7 7.0
Qwen3-VL-8B (CALIBRA) 8.0 7.8 7.6 7.5 7.9
Gemini-2.5-Flash (CALIBRA) 8.9 8.7 8.5 8.6 8.8
4.1s
Avg. Latency
routed mode
8.7s
P95 Latency
routed mode
1.6 GB
Peak RAM
routed mode
93%
Coverage
kept predictions

Deployment efficiency

latency vs coverage
System / Mode Avg. Latency ↓ (s) P95 Latency ↓ (s) Peak RAM ↓ (GB) Coverage ↑ (%)
Deterministic (FFT+Peak) 0.03 0.05 0.35 100
Single-agent (CoT) 2.4 5.1 1.1 98
CALIBRA (Routed mix) 4.1 8.7 1.6 93
CALIBRA (Strict pass@3) 7.8 15.6 1.9 91
Figure 3 Benchmark plot

Segment-level MAE for prompting baselines and CALIBRA across backbones.

Qualitative Results

Figure 4

A representative challenging segment where CALIBRA resolves the final prediction through routing, evidence verification, and calibrated arbitration.

Ablation

The main gains come from routing, critic checks, repair, and calibrated abstention.

Figure 5 Ablation

Component ablation of CALIBRA on the best-performing configuration.

3.42
Full System MAE
best configuration
3.94
w/o Critic
largest drop
3.82
w/o Repair
higher error
3.66
w/o Routing
less efficient selection

BibTeX

@InProceedings{Sakib_2026_CVPR,
    author    = {Sakib, Shadman and Shinde, Gaurav and Roy, Nirmalya},
    title     = {CALIBRA: Calibration-Aware Multi-Agent Verification for Contactless Physiological Monitoring},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)},
    year      = {2026}
}