arXiv subjects

Zhiqiang Tang

Publications and source records attributed to Zhiqiang Tang.

22 records · Page 2Linked to original sources

arXiv2018

Quantized Densely Connected U-Nets for Efficient Landmark Localization

In this paper, we propose quantized densely connected U-Nets for efficient visual landmark localization. The idea is that features of the same semantic meanings are globally reused across the stacked U-Nets. This dense connectivity largely improves the information flow, yielding improved localization accuracy. However, a vanilla dense design would suffer from critical efficiency issue in both training and testing. To solve this problem, we first propose order-K dense connectivity to trim off long-distance shortcuts; then, we use a memory-efficient implementation to significantly boost the training efficiency and investigate an iterative refinement that may slice the model size in half. Finally, to reduce the memory consumption and high precision operations both in training and testing, we further quantize weights, inputs, and gradients of our localization network to low bit-width numbers. We validate our approach in two tasks: human pose estimation and face alignment. The results show that our approach achieves state-of-the-art localization accuracy, but using ~70% fewer parameters, ~98% less model size and saving ~75% training memory compared with other benchmark localizers. The code is available at https://github.com/zhiqiangdon/CU-Net.

cs.CV↗

arXiv2018

Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation

Random data augmentation is a critical technique to avoid overfitting in training deep neural network models. However, data augmentation and network training are usually treated as two isolated processes, limiting the effectiveness of network training. Why not jointly optimize the two? We propose adversarial data augmentation to address this limitation. The main idea is to design an augmentation network (generator) that competes against a target network (discriminator) by generating `hard' augmentation operations online. The augmentation network explores the weaknesses of the target network, while the latter learns from `hard' augmentations to achieve better performance. We also design a reward/penalty strategy for effective joint training. We demonstrate our approach on the problem of human pose estimation and carry out a comprehensive experimental analysis, showing that our method can significantly improve state-of-the-art models without additional data efforts.

cs.CV↗

arXiv2018

Silicon Oxide Electron-Emitting Nanodiodes

Electrically driven on-chip electron sources that do not need to be heated have been long pursued because the current thermionic electron sources show the problems of high power consumption, slow temporal response, bulky size, etc., but their realization remains challenging. Here we show that a nanogap formed by two electrodes on a silicon oxide substrate functions as an electron-emitting nanodiode after the silicon oxide in the nanogap is electrically switched to a high-resistance conducting state. A nanodiode based on graphene electrodes can be turned on by a voltage of ~7 V in ~100 ns and show an emission current of up to several microamperes, corresponding to an emission density of ~10^6 A cm^-2 and emission efficiency as high as 16.6%. We attribute the electron emission to be generated from a metal-insulator-metal tunneling diode on the substrate surface formed by the rupture of conducting filaments in silicon oxide. An array of 100 nanodiodes exhibits a global emission density of 5 A cm^-2 and stable emission with negligible current degradation over tens of hours under modest vacuum. The combined advantages of a low operating voltage, fast temporal response, high emission density and efficiency, convenient fabrication and integration, and stable emission in modest vacuum make silicon oxide electron-emitting nanodiodes a promising on-chip alternative to thermionic emission sources.

physics.app-ph↗

arXiv2018

Toward Marker-free 3D Pose Estimation in Lifting: A Deep Multi-view Solution

Lifting is a common manual material handling task performed in the workplaces. It is considered as one of the main risk factors for Work-related Musculoskeletal Disorders. To improve work place safety, it is necessary to assess musculoskeletal and biomechanical risk exposures associated with these tasks, which requires very accurate 3D pose. Existing approaches mainly utilize marker-based sensors to collect 3D information. However, these methods are usually expensive to setup, time-consuming in process, and sensitive to the surrounding environment. In this study, we propose a multi-view based deep perceptron approach to address aforementioned limitations. Our approach consists of two modules: a "view-specific perceptron" network extracts rich information independently from the image of view, which includes both 2D shape and hierarchical texture information; while a "multi-view integration" network synthesizes information from all available views to predict accurate 3D pose. To fully evaluate our approach, we carried out comprehensive experiments to compare different variants of our design. The results prove that our approach achieves comparable performance with former marker-based methods, i.e. an average error of $14.72 \pm 2.96$ mm on the lifting dataset. The results are also compared with state-of-the-art methods on HumanEva-I dataset, which demonstrates the superior performance of our approach.

cs.CV↗