OpenVSLAM-Based Development Framework for Web-Based VR Tours Using 360VR Videos and Its Extensions

This paper proposes an extended development framework for web-based VR tours using 360VR videos based on OpenVSLAM. The authors have already proposed the development framework for web-based VR tours as navigation contents of 360VR videos and as walkthrough contents of 360VR images. However, it is not easy to prepare 360VR images of different locations for the walkthrough contents if the number of those images is too many. So, the authors have also proposed new development framework for web-based VR tours as both navigation contents and walkthrough contents using 360VR videos, not using 360VR images. For the walkthrough contents, we use 360VR videos because the required 360VR images can be extracted as keyframes of the same scenes from the 360VR videos. Furthermore, when there are many scenes, i.e. many 360VR images in one VR tour, it will become very convenient that the user can go to his/her desired scenes directly by the keyword search. Also, when 360VR videos created as educational contents include audio narrations, it will become also very convenient if the multi-language subtitles are available from the audio narrations. So, the authors extended their development framework for web-based 360VR tours to support the keyword search and multi-language subtitles using Google Cloud Vision AI and free AssemblyAI services.