https://github.com/jesly-joji/violence-detection

Violence Detection
https://github.com/jesly-joji/violence-detection

hybrid-architecture keras streamlit transfer-learining

Last synced: about 2 months ago
JSON representation

Violence Detection

• Host: GitHub
• URL: https://github.com/jesly-joji/violence-detection
• Owner: Jesly-Joji
• Created: 2024-09-08T06:28:42.000Z (almost 2 years ago)
• Default Branch: main
• Last Pushed: 2024-11-19T01:55:50.000Z (over 1 year ago)
• Last Synced: 2025-04-05T14:13:29.051Z (about 1 year ago)
• Topics: hybrid-architecture, keras, streamlit, transfer-learining
• Language: Jupyter Notebook
• Homepage: https://violence-detection.streamlit.app
• Size: 20.6 MB
• Stars: 1
• Watchers: 1
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Violence Detection

Used the hybrid CNN-LSTM Architecture for the video classification


Used MobileNetV2 for CNN as it is Computationally efficient


TOOLS: Numpy, Matplotlib, Seaborn, Keras, Open CV, Streamlit


Live Link:https://violence-detection.streamlit.app


 F1-SCORE: 91.6%


CONFUSION MATRIX


![736c1f16-4e30-4c55-8023-59ce9f2e5c6b](https://github.com/user-attachments/assets/1461854d-135a-454b-b2f4-169425c26a22)


![e0913910-ee39-47f3-9518-49466e9286d5](https://github.com/user-attachments/assets/bb107b2a-1d00-43f9-9196-2166370c0191)        

![cc96ef78-6a08-48df-aeb4-08b8fe5653bf](https://github.com/user-attachments/assets/58791db6-2304-436d-a699-04e3906f57e0)