Hiroharu Kato

Hiroharu Kato

Researcher

Preferred Networks, Inc.

Biography

I’m a researcher at Preferred Networks, Inc.

Interests

  • Computer Vision
  • Machine Learning
  • Computer Graphics

Education

  • Ph.D. in Information Science and Technology, 2020

    The University of Tokyo

  • M.S. in Information Science and Technology, 2014

    The University of Tokyo

  • B.S. in Engineering, 2012

    The University of Tokyo

Experience

 
 
 
 
 

Researcher

Preferred Networks, Inc.

Dec 2018 – Present
 
 
 
 
 

Graduate Student (Doctoral Course)

The University of Tokyo

Oct 2016 – Sep 2020

Research on single-view 3D object reconstruction.

  • Thesis: “Learning 3D Mesh Reconstruction from 2D Images”
  • Superviser: Prof. Tatsuya Harada
 
 
 
 
 

Software Engineer

Sony Corporation

Apr 2014 – Nov 2018
 
 
 
 
 

Graduate Student (Master’s Course)

The University of Tokyo

Apr 2012 – Mar 2014

Research on image generation via image features.

  • Thesis: “Image Reconstruction from Bag-of-Visual-Words”
  • Superviser: Prof. Tatsuya Harada
 
 
 
 
 

Undergraduate Student

The University of Tokyo

Apr 2008 – Mar 2012

Research on detecting anomalous objects for robot vision.

  • Thesis: “Visual Anomaly Detection from Small Samples for Mobile Robots”
  • Superviser: Prof. Tatsuya Harada

Publications

Monocular Differentiable Rendering for Self-Supervised 3D Object Detection

In European Conference on Computer Vision (ECCV) 2020.
Unsupervised 3D car localization by placing a 3D object at a random 3D location and optimizing it with differentiable rendering.

Differentiable Rendering: A Survey

In arXiv 2020.
The first comprehensive survey on differentiable rendering.

Melody Generation for Pop Music via Word Representation of Musical Properties

In arXiv 2017.
Automatic melody generation for pop music has been a long-time aspiration for both AI researchers and musicians. However, learning to …

Talks

Recent trends in differentiable rendering

In Symposium on Sensing via Image Information (SSII) 2020.

Neural 3D Mesh Renderer

In Annual Conference of the Japanese Society for Artificial Intelligence (JSAI) 2019.

[CVPR 2018] Neural 3D Mesh Renderer

In Meeting on Image Recognition and Understanding (MIRU) 2018.

Modality translation and image generation

In Symposium on Sensing via Image Information (SSII) 2018.

[CVPR 2014] Image Reconstruction from Bag-of-Visual-Words

In Meeting on Image Recognition and Understanding (MIRU) 2014.

Professional Activities

  • Reviewer: CVPR (2019-2020), ICCV (2019), ECCV (2020), SIGGRAPH Asia (2019), TVCG (2019-2020), AAAI (2020), IJCAI (2018)
  • Evaluator: MIRU (2019-2020)