Violence Classification Using Audio Analysis...
summary
Presents a project that builds a labeled audio dataset and trains a neural network to classify violent audio into five severity classes using audio analysis and embeddings. Also applies text embeddings to transcriptions or related phrases to measure semantic closeness to aggression, combining audio and text signals to assign both presence and degree of violence in audio segments.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 3:21Project background
- 7:19Key sound identification
- 10:41Audio data cleaning
- 13:11Model training with spectrograms
- 16:50Text analysis for violence
- 19:24Technical demo
- 23:08Challenges and future work
- 25:56Conclusion and applications
- 26:27Q&A
related talks
44mNeural Networks in Sound Analysis and Generation
28mHow to use Deep Learning to process audio data and classify sounds.
22mImplementation and analysis of deep learning models for speech classification
45mAuditory Intelligence : beyond speech recognition
31mClassifying Audio into Types using Python
25mLooking at Sound
22mDeep learning module recognizes video sounds
30mStarting deep learning projects without enough labelled data