Pith Number

pith:TWQFIPIH

pith:2026:TWQFIPIHAOY5FH5HUOOLI2LYZN

not attested not anchored not stored refs resolved

OmniVL-Guard Pro: A Tool-Augmented Agent for Omnibus Vision-Language Forensics

Jinjie Shen, Lechao Cheng, Nan Pu, Shengeng Tang, Tianrui Hui, Yaxiong Wang, Yuchen Zhang, Yujiao Wu, Zheng Huang, Zhun Zhong

A tool-augmented agent overcomes the limits of self-contained vision-language models in open-world forgery detection by integrating external tools and specialized training.

arxiv:2605.16962 v1 · 2026-05-16 · cs.CV · cs.AI

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{TWQFIPIHAOY5FH5HUOOLI2LYZN}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

OmniVL-Guard Pro achieves state-of-the-art performance across various tasks, and exhibits strong zero-shot generalization by extending unified forensics from closed-world prediction to open-world clues-driven reasoning via tool integration, Tree-Structured Self-Evolving Tool Trajectory Generation, and Checker-Guided Agentic Reinforcement Learning.

C2weakest assumption

That the proposed tool environment (real-time event search, local cropping and zooming, edge-anomaly screening, face detection, video frame extraction, and SAM3-based segmentation) combined with the self-evolving trajectories and process-level RL supervision will overcome the practical ceiling of self-contained MLLMs in dynamic open-world forensics.

C3one line summary

OmniVL-Guard Pro is a tool-augmented agent for open-world vision-language forgery detection that integrates multiple external tools with self-evolving trajectory generation and checker-guided reinforcement learning to achieve claimed SOTA performance and zero-shot generalization.

References

35 extracted · 35 resolved · 0 Pith anchors

[1] Is the event time correctly extracted?

[2] Is the subject correctly identified?

[3] Is the location correctly identified?

[4] Is the key action correctly described?

[5] Correct” or “Incorrect

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-20T00:03:33.168128Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

9da0543d0703b1d29fa7a39cb46978cb62cd9277fa02b51f95d85d9bff3abbea

Aliases

arxiv: 2605.16962 · arxiv_version: 2605.16962v1 · doi: 10.48550/arxiv.2605.16962 · pith_short_12: TWQFIPIHAOY5 · pith_short_16: TWQFIPIHAOY5FH5H · pith_short_8: TWQFIPIH

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/TWQFIPIHAOY5FH5HUOOLI2LYZN \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 9da0543d0703b1d29fa7a39cb46978cb62cd9277fa02b51f95d85d9bff3abbea

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "c75fb3613d43c5dd4d9ebfb8c095f173944cbbb92912d4be0a957dffec110dd0",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2026-05-16T12:26:04Z",
    "title_canon_sha256": "1717575c6981088d892d4cdd0e86c81857b4cb03f7c9b86b68fcb640dcbac8df"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.16962",
    "kind": "arxiv",
    "version": 1
  }
}