ETRI SW콘텐츠연구소 차세대콘텐츠연구본부 및 서울시립대 연구과제

모션 플랫폼에서 발생하는 VR멀미 정량화 및 저감 연구

see requirements.txt

pip install -r requirements.txt

project
│   README.md
│   preprocess_video.py
|   preprocess_motion.py
|   processing.py
|   make_video_with_SSQ.py
|   CalQ.cs
|   PlayMenu.cs
|───data
│   |───raw
│   |   |─motion < from unity synchronized with video, 3hz sample rate
│   |   |─video < from unity synchronized with motion, 3hz sample rate
│   |───preprocessed
│   |   |─motion < preprocess_motion.py; probability of motion in simulation level
│   └───processed
│       |─video < make_video_with_SSQ.py; highlight SSQ and optical flows
│       |─table < processing.py; pending currently
│       |─graph < processing.py < preprocess_video.py
│───old
|───test

1. subsampling Motion of Platform(MoP)
   Raw data of MoP which is sampled as 3hz
   → Making lookup-table by scanning all motion in simulation(a video; E.g. probably there are 2 MoP in S1_pitch, up and down)
   → At frame level, take probability of the frame's probability with the lookup table
   → Save probability as json; Motion Probability as json,
2. Subsampling optical flow
   Raw data of optical flow which is sampled as 3hz
   → Polarization all optical flow
   → Making lookup-table by scanning all optical flows in simulation(a video) at frames and all pixels in frames(pixels in frame*frames in simulation; E.g. 1024*768*60 at 20s simulation)
   → At frame level, take probability of the polarized pixel's optical flow with the lookup table
   → Store Optical Probability in memory.
3. Calculate KLD and entropy of optical flow
   With Motion Probability as json and Optical Probability in memory
   → Make KLD at a frame and entropy of optical flow at a frame
   → Save it as figure(.jpg)

• preprocess_motion.py

Subsampling(3hz) raw motion data of txt file to probability of all motion vector in experiment as json and excel file

• preprocess_video.py --FIXME--

Subsampling(3hz) raw video, however 1024*768 pixel probability are hard to save as json or other format.For now, it used to import optical flow data in processing.py.

• processing.py --FIXME--

From subsampled data by preprocess_video.py and preprocess_motion.py, this module calculate KLD and Entropy of optical flow and save it as jpg.

• make_video_with_SSQ.py

To analysis when VR sickness at simulation, it visualize optical flow of video and highlight SSQ

• process raw data of motion of platform to probability density

python preprocecess_motion.py

• process raw data of video and preprocessed MoP to KLD and Entropy of Optical Flow

python processing.py

• comment project
• add calculate probability of optical flow in video at frame level and export data
• modulize processing.py