Yao Yanxin

Yao Yanxin


Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 November 2019 in Electronics
Verified
Journal Article
Open access
PDF file
TSTingzhu Sun
Tingzhu Sun
Weidong FangWeidong Fang
Weidong Fang
Wei ChenWei Chen
Wei Chen
YYYao Yanxin
Yao Yanxin
FBFangming Bi
Fangming Bi
BWBaolei Wu
Baolei Wu
Although image inpainting based on the generated adversarial network (GAN) has made great breakthroughs in accuracy and speed in recent years, they can only process low-resolution images because of memory limitations and difficulty in training. For high-resolution images, the inpainted regions become blurred and the unpleasant boundaries become visible. Based on the current advanced image generation network, we proposed a novel high-resolution image inpainting method based on multi-scale neural network. This method is a two-stage network including content reconstruction and texture detail restoration. After holding the visually believable fuzzy texture, we further restore the finer details to produce a smoother, clearer, and more coherent inpainting result. Then we propose a special application scene of image inpainting, that is, to delete the redundant pedestrians in the image and ensure the reality of background restoration. It involves pedestrian detection, identifying redundant pedestrians and filling in them with the seemingly correct content. To improve the accuracy of image inpainting in the application scene, we proposed a new mask dataset, which collected the characters in COCO dataset as a mask. Finally, we evaluated our method on COCO and VOC dataset. the experimental results show that our method can produce clearer and more coherent inpainting results, especially for high-resolution images, and the proposed mask dataset can produce better inpainting results in the special application scene.

28 June 2021 in IEEE Access
Verified
Journal Article
Open access
YYYao Yanxin
Yao Yanxin
ZNZhengwei Ni
Zhengwei Ni
WHWanqiu Hu
Wanqiu Hu
Mehul MotaniMehul Motani
Mehul Motani
Energy harvesting (EH) relay communication systems with decoding energy costs in multiple block cases have not been widely studied. This paper investigates the relay network with a decode-and-forward relay powered by EH. Unlike other works, we consider the relay with energy decoding costs which harvests random energy from both a dedicated transmitter and other ambient radio-frequency (RF) sources. The EH relay adopts a harvest-receive-forward time-switching architecture. We optimize the time fractions of the three phases and the reception rate at the relay to maximize the offline throughput for single and multiple block cases under two EH scenarios. The multi-block optimization problem constitutes a complex non-convex problem, which we decouple into a single block problem with two auxiliary variables determined by an outer optimization problem. The original problem is finally solved at the cost of linear optimization after series of tricks. Several conclusions are derived: (i) energy storage is necessary (unnecessary) when the relay harvests energy from the transmitter (ambient RF sources), (ii) the optimal reception rate remains unchanged, while the optimal time fractions vary with the energy harvested from ambient RF sources leading to different average throughput. We give numerical simulations to verify our theoretical analysis.

2 January 2023 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
YYYao Yanxin
Yao Yanxin
YCYibo Chen
Yibo Chen
HYHan Yao
Han Yao
ZNZhengwei Ni
Zhengwei Ni
Mehul MotaniMehul Motani
Mehul Motani
Most approaches to resource allocation in energy harvesting systems for Internet-of-Things (IoT) networks do not consider real-time periodic task allocation with Quality of Service (QoS). This paper studies the offline two-dimensional optimization problem for allocating randomly harvested energy flowing causally between slots into different real-time periodic tasks on an IoT device. We formulate an energy allocation problem, for tasks with different energy costs and requested QoS, which aims to maximize a convex reward function subject to energy causality (EC), energy saturation (ES) and task executable (TE) constraints within a certain length of time. We decouple the optimization problem into two subproblems after analysis. First, we propose a novel method to allocate energy to slots based on the Karush-Kuhn-Tucher conditions only considering the EC & ES constraints. The proposed method outperforms the state of the art “Tunnel Policy”, based on geometric programming. Next, an adaption is made to satisfy the TE constraints by allocating in the original constant power slots directly without iteratively checking the wasted energy. Finally, the energy already allocated to slots is put into tasks to complete the two-dimensional allocation. The effectiveness of the proposed methods and task framework are validated by extensive experiments.

23 May 2020
Verified
Book Chapter
Closed access
PDF file
CWChenxu Wang
Chenxu Wang
YYYao Yanxin
Yao Yanxin
ZNZhengwei Ni
Zhengwei Ni
RBRajshekhar V. Bhat
Rajshekhar V. Bhat
MMMehul Motani
Mehul Motani