AIbase
EN
Home
Vivit B 16x2 Kinetics400 UCF Crime Finetuned AbnormalVideosOnly
V

Vivit B 16x2 Kinetics400 UCF Crime Finetuned AbnormalVideosOnly

Developed by Prabesh06
This model is a video classification model based on the ViViT architecture, specifically fine-tuned for anomaly video detection tasks
Video Processing
Transformers
Open Source License:MIT#Video Anomaly Detection#Violence Recognition#Surveillance Video Analysis
Downloads 15
Release Time : 1/11/2025

Model Overview

A video understanding model based on the ViViT architecture, fine-tuned on the UCF-Crime dataset, focusing on abnormal behavior detection

Model Features

Video Anomaly Detection
Specifically designed to detect and classify abnormal behaviors in videos
Transformer-Based Architecture
Utilizes the ViViT architecture, combining the advantages of Vision Transformers in video understanding
Pre-training and Fine-tuning Paradigm
Pre-trained on Kinetics400 and fine-tuned on the UCF-Crime dataset

Model Capabilities

Video Classification
Anomaly Detection
Spatio-Temporal Feature Extraction

Use Cases

Security Surveillance
Surveillance Video Anomaly Detection
Automatically detects abnormal behaviors or events in surveillance videos
Achieved an accuracy of 9.62% on the evaluation set
Video Analysis
Video Content Understanding
Analyzes video content and identifies abnormal scenes
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