INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
Scenery Soundscaping using Cross-Modal Information Retrieval
Authors Name:
Abheetha Pradhan
, Vikram C , Dr. Manjula K
Unique Id:
IJSDR2207067
Published In:
Volume 7 Issue 7, July-2022
Abstract:
Human beings associate certain sceneries with certain ambient sounds. A picture of the ocean can make us think of the sound of ocean waves, or a picture of a busy market can make us imagine the sounds of people talking and the sound of vehicles. But computers do not have this innate ability to associate audio to a visual image. If you separate the audio from its corresponding video, the computer would not be able to associate these with each other, unless programmed to do so explicitly. Cross-Modal Information Retrieval (CMIR) is a technique that measures similarities between different types of data, usually media, which here would be audio in the form of mp3 files, and images in the form of jpg files. The Scenery Soundscaping System accurately predicts the ambient sounds that you would hear in the scenery of a particular image using this technique. The project also creates a web app for the user to provide image inputs to the Scenery Soundscaping System and to listen to the soundscape predicted by it.
Keywords:
Deep Learning, Cross-Modal Information Retrieval
Cite Article:
"Scenery Soundscaping using Cross-Modal Information Retrieval", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.7, Issue 7, page no.483 - 487, July-2022, Available :http://www.ijsdr.org/papers/IJSDR2207067.pdf
Downloads:
000337211
Publication Details:
Published Paper ID: IJSDR2207067
Registration ID:201016
Published In: Volume 7 Issue 7, July-2022
DOI (Digital Object Identifier):
Page No: 483 - 487
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn