IP-Bench: Benchmark for Image Protection Methods in Image-to-Video Generation Scenarios

Dimensions overall score 4.0

• I2VGuard demonstrates notable robustness against preprocessing attacks, maintaining a degradation rate above 0.02 across most models.result 85%
  Evidencepartial

  I2VGuard [13] demonstrates notable robustness against pre- processing attacks, maintaining a degradation rate above 0.02 across most models.

  Implicationpartial

  This is a specific result presented in the paper with a quantitative measure.

  Verificationpartial

  partial

• VGMShield demonstrates strong performance on specific models such as W AN, achieving a degradation rate of 0.036.result 85%
  Evidencepartial

  VGMShield [14] demonstrates strong performance on specific models such as W AN [22], achiev- ing a degradation rate of 0.036.

  Implicationpartial

  This is a specific result presented in the paper with a quantitative measure for a particular model.

  Verificationpartial

  partial

• PhotoGuard achieves a balanced performance with a moderate PSNR of 31.77 and a degradation rate of 0.041 on Skyreel.result 85%
  Evidencepartial

  PhotoGuard [15] achieves a balanced performance, with a moderate PSNR of 31.77 and a degradation rate of 0.041 on Skyreel [23].

  Implicationpartial

  This is a specific result presented in the paper with quantitative measures for a particular model.

  Verificationpartial

  partial

• No existing image protection method transfers effectively across different architectures.limitation 80%
  Evidencepartial

  Furthermore, no existing method transfers effectively across different archi- tectures

  Implicationpartial

  This is a stated limitation and a motivation for the benchmark, implying a finding from prior work.

  Verificationpartial

  partial

• IP-Bench evaluates protection methods across 5 representative I2V models featuring distinct structures, such as DiT and U-Net.method 95%
  Evidencepartial

  I2I-based approaches) across 5 widely used I2V models fea- turing distinct structures, such as DiT and U-Net [19, 20].

  Implicationpartial

  This is explicitly stated in the abstract and supported by the description of the I2V models used.

  Verificationpartial

  partial

• IP-Bench is the first systematic benchmark designed to evaluate image protection methods in image-to-video (I2V) generation scenarios.method 95%
  Evidencepartial

  To address this challenge, we propose IP-Bench (Image Protection Bench), the first systematic benchmark designed to evaluate protection methods in I2V generation scenarios.

  Implicationpartial

  This is explicitly stated in the abstract and reinforced throughout the introduction.

  Verificationpartial

  partial

• IP-Bench evaluates 6 representative protection methods and 5 state-of-the-art I2V models.method 95%
  Evidencepartial

  This benchmark examines 6 representative protection methods and 5 state-of-the-art I2V models.

  Implicationpartial

  This is explicitly stated in the abstract and supported by the overview tables in the paper.

  Verificationpartial

  partial

• IP-Bench systematically evaluates protection methods' robustness against two robustness attack strategies under practical scenarios.method 95%
  Evidencepartial

  Furthermore, our work systematically evaluates protection methods' robustness with two robustness attack strategies under practical scenarios and analyzes their cross-model & cross-modality transferability.

  Implicationpartial

  This is explicitly stated in the abstract and detailed in the methodology section.

  Verificationpartial

  partial

Competitive landscape

A benchmark for evaluating image protection methods in image-to-video generation to combat misuse.