ForensicZip: More Tokens are Better but Not Necessary in Forensic Vision-Language Models

• URL: http://arxiv.org/abs/2603.12208v1
• Date: Thu, 12 Mar 2026 17:30:49 GMT
• Title: ForensicZip: More Tokens are Better but Not Necessary in Forensic Vision-Language Models
• Abstract summary: We introduce ForensicZip, a training-free framework that reformulates token compression from a forgery-driven perspective.<n>ForensicZip achieves $2.97times$ speedup and over 90% FLOPs reduction while maintaining state-of-the-art detection performance.
• Score: 108.5042835056188
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multimodal Large Language Models (MLLMs) enable interpretable multimedia forensics by generating textual rationales for forgery detection. However, processing dense visual sequences incurs high computational costs, particularly for high-resolution images and videos. Visual token pruning is a practical acceleration strategy, yet existing methods are largely semantic-driven, retaining salient objects while discarding background regions where manipulation traces such as high-frequency anomalies and temporal jitters often reside. To address this issue, we introduce ForensicZip, a training-free framework that reformulates token compression from a forgery-driven perspective. ForensicZip models temporal token evolution as a Birth-Death Optimal Transport problem with a slack dummy node, quantifying physical discontinuities indicating transient generative artifacts. The forensic scoring further integrates transport-based novelty with high-frequency priors to separate forensic evidence from semantic content under large-ratio compression. Experiments on deepfake and AIGC benchmarks show that at 10\% token retention, ForensicZip achieves $2.97\times$ speedup and over 90\% FLOPs reduction while maintaining state-of-the-art detection performance.

Related papers

• TIGA: Trajectory-Injected Generative Attack against Black-box AIGC Detectors [76.4118981404995]
  Trajectory-Injected Generative Attack (TIGA) is a source-image-free and training free framework that generates detector-evasive images.<n>TIGA steers the latent Denoising Diffusion Implicit Model (DDIM) trajectory so that adversarial properties emerge during generation.<n>TIGA achieves strong blackbox attack performance, transferability, and high robustness under common post-processing operations.
  arXiv  Detail & Related papers  (2026-07-28T15:51:39Z)
• EVAS: Efficient Multimodal Temporal Forgery Localization via Audio-Visual Synergy and Steered Boundary Calibration [14.088182834458843]
  We propose EVAS, an end-to-end framework for temporal forgery localization.<n>We introduce a Boundary-Aware Refinement strategy to achieve steered boundary calibration.<n> EVAS achieves state-of-the-art average localization accuracy and average recall across three benchmark datasets.
  arXiv  Detail & Related papers  (2026-07-05T19:33:21Z)
• ForensicsTok: Forensics-Guided Tokenized Modeling for Image Tampering Localization [31.83661814822312]
  We propose ForensicsTok, which reformulates image manipulation localization as an autoregressive sequence generation task.<n>ForensicsTok directly generates spatially grounded token sequences, enabling precise mask prediction without intermediary supervision.
  arXiv  Detail & Related papers  (2026-06-23T13:06:18Z)
• Deep Residual Injection for Full-Spectrum Forensic Signal Perception in Multimodal Large Language Models [51.099179436350376]
  Multimodal large language models (MLLMs) have been increasingly adopted in forensics for their robust semantic understanding.<n>We propose Deep Visual Residual MLLM (Deep-VRM) to preserve early semantic processing while injecting artifact-specific visual signals as a residual path into an intermediate layer.<n>This enables later layers to jointly model semantic reasoning and signal-level forensic cues, and surprisingly, the model learns to adaptively leverage different levels of forensic signals depending on the input.
  arXiv  Detail & Related papers  (2026-06-14T16:05:45Z)
• FastOCR: Dynamic Visual Fixation via KV Cache Pruning for Efficient Document Parsing [51.905216364362325]
  We propose FastOCR, a training-free framework with two complementary modules.<n>FastOCR retains 98% of the unpruned model's accuracy while attending to only 5% of the visual tokens per decoding step.
  arXiv  Detail & Related papers  (2026-05-17T13:39:47Z)
• Attribution-Guided Multimodal Deepfake Detection via Cross-Modal Forensic Fingerprints [4.598350938853635]
  We argue that a detector incapable of identifying how a video was forged is likely learning the wrong signal.<n>Unlike binary detection, attribution-guided learning imposes a stronger geometric constraint on the shared embedding space.<n>We propose the Attribution-Guided Multimodal Deepfake Detection framework, which jointly learns to detect and attribute manipulation.
  arXiv  Detail & Related papers  (2026-04-29T09:11:13Z)
• RetentiveKV: State-Space Memory for Uncertainty-Aware Multimodal KV Cache Eviction [3.2907393353763243]
  Existing KV cache compression methods rely on the "persistence of importance" hypothesis to prune tokens.<n>We propose RetentiveKV, an entropy-driven KV cache optimization method that reformulates KV eviction from "discrete context truncation" to "continuous memory evolution" based on State Space Models.
  arXiv  Detail & Related papers  (2026-04-14T08:17:53Z)
• Word-Anchored Temporal Forgery Localization [6.691985085293349]
  We propose word-anchored temporal forgery localization (WAFL) to derive candidate forgery proposals.<n>We first analyze the essence of temporal forgeries and identify the minimum meaningful forgery units, word tokens, and then align data preprocessing with the natural linguistic boundaries of speech.<n>To overcome the extreme class imbalance inherent to forgery detection, we design the artifact-centric asymmetric (ACA) loss.
  arXiv  Detail & Related papers  (2026-03-06T12:33:28Z)
• ForgeryVCR: Visual-Centric Reasoning via Efficient Forensic Tools in MLLMs for Image Forgery Detection and Localization [62.03035862528452]
  ForgeryVCR is a framework that materializes imperceptible traces into explicit visual intermediates via Visual-Centric Reasoning.<n>ForgeryVCR achieves state-of-the-art (SOTA) performance in both detection and localization tasks.
  arXiv  Detail & Related papers  (2026-02-15T11:14:47Z)
• Detecting Overflow in Compressed Token Representations for Retrieval-Augmented Generation [49.48204107529758]
  We define token overflow as a regime in which compressed representations no longer contain sufficient information to answer a given query.<n>In this paper, we find that query-agnostic saturation statistics reliably separate compressed from uncompressed token representations.<n>Lightweight probing classifiers over both query and context xRAG representations detect overflow with 0.72 AUC-ROC on average.<n>These results advance from query-independent diagnostics to query-aware detectors, enabling low-cost pre-LLM gating to mitigate compression-induced errors.
  arXiv  Detail & Related papers  (2026-02-12T18:15:08Z)
• HIPPO: Accelerating Video Large Language Models Inference via Holistic-aware Parallel Speculative Decoding [48.55833840968632]
  Speculative decoding has emerged as a promising approach to accelerate LLM inference without sacrificing output quality.<n>We propose HIPPO, a holistic-aware parallel speculative decoding framework.<n> experiments on four video-LLMs across six benchmarks demonstrate HIPPO's effectiveness, yielding up to 3.51x speedup.
  arXiv  Detail & Related papers  (2026-01-13T07:02:43Z)
• Towards Robust DeepFake Detection under Unstable Face Sequences: Adaptive Sparse Graph Embedding with Order-Free Representation and Explicit Laplacian Spectral Prior [12.202765237400143]
  We propose a Laplacian-Regularized Graph Convolutional Network (LR-GCN) that robustly detects DeepFakes from noisy or unordered face sequences.<n>LR-GCN achieves state-of-the-art performance and significantly improved robustness under severe global and local disruptions.
  arXiv  Detail & Related papers  (2025-12-08T12:31:07Z)
• SODiff: Semantic-Oriented Diffusion Model for JPEG Compression Artifacts Removal [50.90827365790281]
  SODiff is a semantic-oriented one-step diffusion model for JPEG artifacts removal.<n>Our core idea is that effective restoration hinges on providing semantic-oriented guidance to the pre-trained diffusion model.<n>SAIPE extracts rich features from low-quality (LQ) images and projects them into an embedding space semantically aligned with that of the text encoder.
  arXiv  Detail & Related papers  (2025-08-10T13:48:07Z)
• Learning Spatiotemporal Inconsistency via Thumbnail Layout for Face Deepfake Detection [41.35861722481721]
  Deepfake threats to society and cybersecurity have provoked significant public apprehension. This paper introduces an elegantly simple yet effective strategy named Thumbnail Layout (TALL) TALL transforms a video clip into a pre-defined layout to realize the preservation of spatial and temporal dependencies.
  arXiv  Detail & Related papers  (2024-03-15T12:48:44Z)
• Towards General Visual-Linguistic Face Forgery Detection [95.73987327101143]
  Deepfakes are realistic face manipulations that can pose serious threats to security, privacy, and trust. Existing methods mostly treat this task as binary classification, which uses digital labels or mask signals to train the detection model. We propose a novel paradigm named Visual-Linguistic Face Forgery Detection(VLFFD), which uses fine-grained sentence-level prompts as the annotation.
  arXiv  Detail & Related papers  (2023-07-31T10:22:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.