sh (김소형 논문) [논문 통과를 진심으로 축하 합니다. 2019년 9월]

keywords

• novel view synthesis
• depth completion
• stereo images
• GAN
• auto encoder, auto decoder

이영학 정윤주 심재창

• We would like to make a best 3D images using two stereo images
• Simple
• Efficient

Meeting Schedule

S.H will visit Andong National Univ. on 6.22.2019 <Welcome!!!>

• 8:00 Daegu, Dr. Lee will pick her up.
• 9:30 Andong National Univ., 406 Eng. Bld II, Dr. Shim's office
• 9:40 Introduction
• 9:40~10:20 SH will present her idea
• 10:20~11:20 Discussion
• 11:20~12:00 Conclusion

References

• NVS Machines
  • Paper in PDF NVS Machines: Learning Novel View Synthesis with Fine-grained View Control
  • paper in web
  • video NVS Machines: Learning Novel View Synthesis with Fine-grained View Control *data

참고 동영상

• Stereo Magnification

  • youtube SIGGRAPH 2018, Stereo Magnification: Learning view synthesis using multiplane images
  • source code in github
  • youtube Stereo Magnification: Learning view synthesis using multiplane images

• References

  • Youtube, Learning-Based View Synthesis for Light Field Cameras (SIGGRAPH Asia 2016)
  • Paper, Source, Learning-Based View Synthesis for Light Field Cameras (SIGGRAPH Asia 2016)

• Etc..

  • AI Learns Video Frame Interpolation
  • NVS Machines: Learning Novel View Synthesis with Fine-grained View Control
  • DeepView: View Synthesis with Learned Gradient Descent, 2018
  • Soft 3D Reconstruction for View Synthesis

• Lesson

  • Lesson, Youtube 25분, 2014: 3d reconstruction with stereo cameras

• Stereo Camera

• [2D-to-Stereo Panorama Conversion Using GAN and Concentric Mosaics]

• [Virtual View Generation Based on 3D-Dense-Attentive GAN Networks]

• [https://youtu.be/3AIpPlzM_qs]

• [https://arxiv.org/abs/1803.06657]