AIbase
EN
Home
Videomae Base Finetuned Ssv2 Finetuned Ucfcrime Ep10
V

Videomae Base Finetuned Ssv2 Finetuned Ucfcrime Ep10

Developed by Tejdeep
Video understanding model based on VideoMAE architecture, pre-trained on SSv2 dataset and fine-tuned for UCF-Crime dataset
Video Processing
Transformers
#Video Anomaly Detection#Crime Recognition#Fine-tuned Model
Downloads 243
Release Time : 7/10/2023

Model Overview

This model is specialized for video understanding, focusing on video content analysis, potentially applicable to security scenarios such as abnormal behavior detection

Model Features

Video Feature Extraction
Based on masked autoencoder pre-training, effectively capturing spatiotemporal features in videos
Anomaly Detection Optimization
Fine-tuned on UCF-Crime dataset, optimized for security scenarios
Efficient Training
Utilizes linear learning rate scheduling and warm-up strategies for higher training efficiency

Model Capabilities

Video Content Understanding
Abnormal Behavior Recognition
Spatiotemporal Feature Extraction

Use Cases

Smart Security
Violence Detection
Identifying violent behaviors in surveillance videos
Evaluation accuracy 80.04%
Abnormal Event Alert
Real-time detection of abnormal events in public areas
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase