Topology matching for fully automatic similarity estimation of 3D shapes
- Masaki Hilaga, Y. Shinagawa, T. Komura, T. Kunii
- Computer ScienceInternational Conference on Computer Graphics and…
- 1 August 2001
A novel technique is proposed, called Topology Matching, in which similarity between polyhedral models is quickly, accurately, and automatically calculated by comparing Multiresolutional Reeb Graphs (MRGs), which operates well as a search key for 3D shape data sets.
Topological Modeling for Visualization
- T. Kunii, H. Nishida, Masaki Hilaga
- Mathematics
- 5 September 1997
Part 1 Foundation: curves the notion of a Riemannian metric local theory of surfaces the classification of surfaces abstract manifolds critical points and Morse theory analyzing human body motions…
Image and video processing on mobile devices: a survey
- Chamin Morikawa, Michihiro Kobayashi, Masaki Hilaga
- Computer ScienceThe Visual Computer
- 21 June 2021
This paper surveys mobile image processing and computer vision applications while highlighting constraints and explaining how the algorithms have been modified/adapted to meet accuracy and performance demands.
High Impact Business Modeling through Smart Cloud Computing
- T. Kunii, Masaki Hilaga, Takeshi Miura
- Computer ScienceInternational Conference on Cyberworlds
- 25 September 2012
Extreme diversities, versatility and dynamism of cloud computing applications with its high availability through smart devices such as smart phones and media are shown to be supported by the cyber world modeling in an incrementally modular abstraction hierarchy (IMAH) as a transparent and highly universal modeling of smart cloud computing businesses to overcome combinatorial explosion of smart computing.
Applications of Topological Modeling and Visualization
- T. Kunii, H. Nishida, Masaki Hilaga
- Computer Science
- 1 December 2012
Mathematics for Topological Modeling
- T. Kunii, H. Nishida, Masaki Hilaga
- Mathematics
- 1 September 2012
Homotopy Type Theory for Big Data
- T. Kunii, Masaki Hilaga
- Computer ScienceInternational Conference on Cyberworlds
- 7 October 2015
This research first presents a generic theoretical framework as an incrementally modular abstraction hierarchy, based on homotopy type theory, provides an axiomatic approach to theorize the potentials of big data in cyber worlds, and shows that the incrementally Modular abstraction hierarchy automate big data application development and eliminates the need for design verification and validation.
SoftNeuro: Fast Deep Inference using Multi-platform Optimization
- Masaki Hilaga, Yasuhiro Kuroda, Yusuke Kozawa
- Computer ScienceArXiv
- 12 October 2021
This work proposes SoftNeuro, a novel, high-performance inference framework with efficient performance tuning that maximizes the inference performance by profiling various routines for each layer and selecting the fastest path.