Search ETDs:
Deep-learning Approaches to Object Recognition from 3D Data
Chen, Zhiang

2017, Master of Sciences, Case Western Reserve University, EMC - Mechanical Engineering.
This thesis focuses on deep-learning approaches to recognition and pose estimation of graspable objects using depth information. Recognition and orientation detection from depth-only data is encoded by a carefully designed 2D descriptor from 3D point clouds. Deep-learning approaches are explored from two main directions: supervised learning and semi-supervised learning. The disadvantages of supervised learning approaches drive the exploration of unsupervised pretraining. By learning good representations embedded in early layers, subsequent layers can be trained faster and with better performance. An understanding of learning processes from a probabilistic perspective is concluded, and it paves the way for developing networks based on Bayesian models, including Variational Auto-Encoders. Exploitation of knowledge transfer--re-using parameters learned from alternative training data--is shown to be effective in the present application.
Wyatt Newman, PhD (Advisor)
M. Cenk Çavusoglu, PhD (Committee Member)
Roger Quinn, PhD (Committee Member)
92 p.

Recommended Citations

Hide/Show APA Citation

Chen, Z. (2017). Deep-learning Approaches to Object Recognition from 3D Data. (Electronic Thesis or Dissertation). Retrieved from https://etd.ohiolink.edu/

Hide/Show MLA Citation

Chen, Zhiang. "Deep-learning Approaches to Object Recognition from 3D Data." Electronic Thesis or Dissertation. Case Western Reserve University, 2017. OhioLINK Electronic Theses and Dissertations Center. 15 Dec 2017.

Hide/Show Chicago Citation

Chen, Zhiang "Deep-learning Approaches to Object Recognition from 3D Data." Electronic Thesis or Dissertation. Case Western Reserve University, 2017. https://etd.ohiolink.edu/

Files

chen_thesis.pdf (3.98 MB) View|Download