Several methodologies published in the literature can be used to construct realistic pore networks for simple rocks, whereas in complex pore geometry formations, as formed in tight reservoirs, such a construction still remains a challenge. A basic understanding of pore structure and topology is essential to overcome the challenges associated with the pore scale modeling of tight porous media. A...

Our goal is to classify 3D models directly using convolutional neural network. Most of existing approaches rely on a set of human-engineered features. We use 3D convolutional neural network to let the network learn the features over 3D space to minimize classification error. We trained and tested over ShapeNet dataset with data augmentation by applying random transformations. We made various vi...

In this paper, we investigate how to design energy-efficient localized routing in a large-scale three-dimensional (3D) wireless network. Several 3D localized routing protocols were proposed to seek either energy efficiency or delivery guarantee in 3D wireless networks. However, recent results [1, 2] showed that there is no deterministic localized routing algorithm that guarantees either deliver...

3D hand pose estimation from single depth image is an important and challenging problem for human-computer interaction. Recently deep convolutional networks (ConvNet) with sophisticated design have been employed to address it, but the improvement over traditional random forest based methods is not so apparent. To exploit the good practice and promote the performance for hand pose estimation, we...

Decentralized control of mobile robotic sensor networks is a fundamental problem in robotics that has attracted intensive research in recent decades. Most of the existing works dealt with two-dimensional spaces. This report is concerned with the problem of decentralized self-deployment of mobile robotic sensor networks for coverage, search, and formation building in three-dimensional environmen...