SafeWatch: An Efficient Safety-Policy Following Video Guardrail Model with Transparent Explanations

🔥🔥 Recent news please check Project page !

An ovewview of the model architecture of SafeWatch-8B. Regarding video input (left), SafeWatch leverages a segmentation model to process the input video into clips based on unsafe events. Then, it samples frames from each event and encodes them into patch tokens. Regarding safety guidelines (right), SafeWatch encodes each policy in parallel with the equivalent RoPE embedding to ensure they are treated with equal importance. Then, for each policy, SafeWatch calculates the relevance score based on its cross-attention with the video tokens and then activates Top-k most informative tokens and prunes the rest. Finally, these tokens are concatenated with the query for decoding.

Performance comparison of SafeWatch-8B with various video guardrail baselines on SafeWatch-Bench-Real. We report the individual accuracy for each category, along with average accuracy (ACC) and F1 Score across all categories. AUPRC is calculated over binary guardrail outputs. Explanations are rated on a numerical scale of [0,10] by both GPT-4o-as-judge and human evaluators. Inference cost is measured by inference time per video. Best performance is in bold.

Comparison of SafeWatch-8B and GPT-4o across fine-grained scenarios in SafeWatch- Bench. We evaluate the average accuracy per subcategory. Hard Benign refers to challenging benign samples that previous models often misclassify as harmful, resulting in high false positives.

Detailed comparison across different guardrail models on the accuracy of each subcategory in SafeWatch-Bench, five existing datasets, i.e., LSPD, XD-V, UCF, FakeSV, FVC, and four new policy categories, i.e., child safety, firearms, accidents.