Zero-Shot Depth from Defocus

Zero-Shot Depth from Defocus | Signal Canvas | ScienceToStartup

Page Freshness

Signal Canvas proof surface

Canonical route: /signal-canvas/zero-shot-depth-from-defocus

stale

Proof freshness: stale
Proof status: partial
Display score: 7/10
Last proof check: 2026-03-30
Score updated: 2026-04-02
Score fresh until: 2026-05-02
References: 62
Source count: 4
Coverage: 83%

This page is showing the last landed evidence receipt and score bundle because the latest proof data is outside the freshness window.

Agent Handoff

Canonical ID zero-shot-depth-from-defocus | Route /signal-canvas/zero-shot-depth-from-defocus

REST example

curl https://sciencetostartup.com/api/v1/agent-handoff/signal-canvas/zero-shot-depth-from-defocus

MCP example

{
  "tool": "search_signal_canvas",
  "arguments": {
    "mode": "paper",
    "paper_ref": "zero-shot-depth-from-defocus",
    "query_text": "Summarize Zero-Shot Depth from Defocus"
  }
}

source_context

{
  "surface": "signal_canvas",
  "mode": "paper",
  "query": "Zero-Shot Depth from Defocus",
  "normalized_query": "2603.26658",
  "route": "/signal-canvas/zero-shot-depth-from-defocus",
  "paper_ref": "zero-shot-depth-from-defocus",
  "topic_slug": null,
  "benchmark_ref": null,
  "dataset_ref": null
}

Evidence Receipt

Route status: building

Claims: 12

References: 62

Proof: Verification pending

Freshness state: computing

Source paper: Zero-Shot Depth from Defocus

PDF: https://arxiv.org/pdf/2603.26658v1

Repository: https://github.com/princeton-vl/FOSSA

Source count: 4

Coverage: 83%

Last proof check: 2026-03-30T20:30:26.682Z

Signal Canvas receipt window

Ready for execution: Zero-Shot Depth from Defocus

/buildability/zero-shot-depth-from-defocus

Build Nowready

Subject: Zero-Shot Depth from Defocus

Verdict

Build Now

Verdict is Build Now because viability and implementation proof cleared the Wave 1 scaffold thresholds.

Time to first demo

Insufficient data

No first-demo timestamp, owner estimate, or elapsed demo receipt is attached to this surface.

Compute envelope

Structured compute envelope

Insufficient data

No data, compute, hardware, memory, latency, dependency, or serving requirement receipt is attached.

Evidence ids

GitHub Code Pulse

Cached

Stars

Health

Last commit

4/5/2026

Forks

Open repository

Claim map

Strong 12Mixed 0Weak 0

Evidencepartial
ZEDD has8.3×more scenes, higher quality ground-truth, higher resolution images, and realistic defocus effect under multiple f-numbers.
Implicationpartial
Directly stated in the abstract and supported by a comparison table.
Verificationpartial
partial
Evidencepartial
The key contribution is a stack attention layer with a focus distance embedding, allowing efficient information exchange across the focus stack.
Implicationpartial
The abstract and description of the network architecture clearly outline this novel component.
Verificationpartial
partial
Evidencepartial
Finally, we develop a new training data pipeline allowing us to utilize existing large-scale RGBD datasets to generate synthetic focus stacks.
Implicationpartial
Stated in the abstract as a key development for training.
Verificationpartial
partial
Evidencepartial
Experiment results on ZEDD and other benchmarks show a significant improvement over the baselines, reducing errors by up to 55.7%.
Implicationpartial
Directly stated in the abstract with a specific quantitative result.
Verificationpartial
partial
Evidencepartial
It reduces AbsRel by 55.7% compared to the second-best method DepthPro [3] on ZEDD, and MSE by 40.4% on DDFF compared to previous state-of-the-art DualFocus [49].
Implicationpartial
Specific quantitative result comparing FOSSA to a baseline on a key metric and dataset.
Verificationpartial
partial
Evidencepartial
We also demonstrate that FOSSA is robust to various factors, including the aperture size and the focus stack size.
Implicationpartial
Stated as a demonstrated capability of the FOSSA method.
Verificationpartial
partial
Evidencepartial
Ours (ViT-B) 0.505 0.824 0.918 0.089 0.420 0.820 0.936 0.091
Implicationpartial
Directly reported in Table 2 with specific model variant and metric.
Verificationpartial
partial
Evidencepartial
ZEDD has8.3×more scenes, higher quality ground-truth, higher resolution images, and realistic defocus effect under multiple f-numbers.
Implicationpartial
This claim is directly stated in the abstract and supported by the comparison in Table 1.
Verificationpartial
partial
Evidencepartial
The key contribution is a stack attention layer with a focus distance embedding, allowing efficient information exchange across the focus stack.
Implicationpartial
This claim is explicitly described in the abstract and further detailed in the description of the network architecture.
Verificationpartial
partial
Evidencepartial
Experiment results on ZEDD and other benchmarks show a significant improvement over the baselines, reducing errors by up to 55.7%.
Implicationpartial
This claim is directly stated in the abstract and supported by the experimental results presented in the text and Table 2.
Verificationpartial
partial
Evidencepartial
It reduces AbsRel by 55.7% compared to the second-best method DepthPro [3] on ZEDD, and MSE by 40.4% on DDFF compared to previous state-of-the-art DualFocus [49].
Implicationpartial
This is a specific quantitative result mentioned in the abstract and elaborated in the text.
Verificationpartial
partial
Evidencepartial
Our method FOSSA outperforms all monocular d
Implicationpartial
This claim is directly supported by the comparison in Table 2, which shows FOSSA achieving better performance metrics than other listed methods.
Verificationpartial
partial

Author intelligence and commercialization panels stay hidden until the proof receipt is verified, cites at least 3 references, includes at least 2 sources, and clears 50% coverage. The paper narrative and citation surfaces remain public while verification is pending.

Zero-Shot Depth from Defocus

Use Signal Canvas as the narrative proof surface

Use this Signal Canvas via API or MCP

Signal Canvas proof surface