← Back to VOLUME 13, ISSUE 11, NOVEMBER 2026
This work is licensed under a Creative Commons Attribution 4.0 International License.
Context-Based Audio-Video Analysis
Prof. S. P. Gade, Dnyaneshwari Mahadev Sawant, Neha Sanjay Patil, Sujal Balasaheb Mane, Payal Yogiraj Sable
👁 2 views📥 2 downloads
Abstract: A video carries two streams of evidence, and each one misses things the other catches. The picture can show a person moving but cannot tell us whether we are hearing footsteps, a siren or rain on a roof. The sound can tell us something happened but often not what produced it. This paper proposes a context-based audio-video analysis system that takes a video as input, separates its audio and visual content, analyzes both, and fuses the results to work out what is going on. The output is a natural-language caption. If the frames show rainfall and the soundtrack contains falling raindrops, for example, the system could produce "It is raining outdoors." The design accounts for background noise, overlapping activities, audio-video synchronization and uneven video quality. Possible uses include surveillance, accessibility, activity recognition and video summarization. The system has not been built yet, so its accuracy, reliability and processing speed still have to be measured.
Keywords: audio-video analysis, multimodal learning, context understanding, activity recognition, audio-visual fusion, event detection, automatic caption generation.
Keywords: audio-video analysis, multimodal learning, context understanding, activity recognition, audio-visual fusion, event detection, automatic caption generation.
How to Cite:
[1] Prof. S. P. Gade, Dnyaneshwari Mahadev Sawant, Neha Sanjay Patil, Sujal Balasaheb Mane, Payal Yogiraj Sable, “Context-Based Audio-Video Analysis,” International Advanced Research Journal in Science, Engineering and Technology (IARJSET), DOI: 10.17148/IARJSET.2026.131104
