Deep ensemble-based hard sample mining for food recognition

Deep neural networks represent a compelling technique to tackle complex real-world problems, but are over-parameterized and often suffer from over- or under-confident estimates. Deep ensembles have shown better parameter estimations and often provide reliable uncertainty estimates that contribute to the robustness of the results. In this work, we propose a new metric to identify samples that are hard to classify. Our metric is defined as coincidence score for deep ensembles which measures the agreement of its individual models. The main hypothesis we rely on is that deep learning algorithms learn the low-loss samples better compared to large-loss samples. In order to compensate for this, we use controlled over-sampling on the identified ”hard” samples using proper data augmentation schemes to enable the models to learn those samples better. We validate the proposed metric using two public food datasets on different backbone architectures and show the improvements compared to the conventional deep neural network training using different performance metrics.     

用迁移率和载流子浓度研究碲镉汞中的深能级

获得了一种研究碲镉汞深能级的方法。通过分析迁移率 载子浓度与温度的关系，可以得到关于深能级的重要依据。     

超宽带雷达人体行为感知研究进展

超宽带 (UWB) 雷达人体行为感知主要研究如何利用人体目标电磁散射回波对位置、行为、意图等进行判别,是光学感知手段的有益补充,应对无光照、地物遮挡、非视距等情况下的应用场合。该文将超宽带雷达人体行为感知研究方法分成基于空间位置和基于微动特征两类技术。在介绍这类技术基本原理的基础上,对比分析了国内外代表性工作的能力现状。最后对超宽带雷达人体行为感知领域的后续重点研究方向进行了展望。     

基于监督对比学习正则化的高分辨率SAR图像建筑物提取方法

近年来,高分辨合成孔径雷达(SAR)图像的智能解译技术在城市规划、变化监测等方面得到了广泛应用.不同于光学图像,SAR图像的获取方式、图像中目标的几何结构等因素制约了现有深度学习方法对SAR图像地物目标的解译效果.该文针对高分辨SAR图像城市区域建筑物提取,提出了基于监督对比学习的正则化方法,其主要思想是增强同一类别像...     

电子辐照深亚微米MOS晶体管的总剂量效应

为了研究电子辐照导致CCD参数退化的损伤机理,以及CCD内不同沟道宽长比的NMOSFET的辐射效应,将与CCD同时流片的两种不同沟道宽长比的深亚微米NMOSFET进行电子辐照实验。分析了电子辐照导致NMOSFET阈值电压和饱和电流退化的情况,以及器件的辐射损伤敏感性。实验结果表明,电子辐照导致两种NMOSFET器件的参数退化情况以及辐射损伤敏感性类似。导致器件参数退化的主要原因是界面陷阱电荷,同时氧化物陷阱电荷表现出了一定的竞争关系。实验结果为研究CCD电子辐照导致的辐射效应提供了基础数据支持。     

Semantic content-based image retrieval: A comprehensive study

The complexity of multimedia contents is significantly increasing in the current digital world. This yields an exigent demand for developing highly effective retrieval systems to satisfy human needs. Recently, extensive research efforts have been presented and conducted in the field of content-based image retrieval (CBIR). The majority of these efforts have been concentrated on reducing the semantic gap that exists between low-level image features represented by digital machines and the profusion of high-level human perception used to perceive images. Based on the growing research in the recent years, this paper provides a comprehensive review on the state-of-the-art in the field of CBIR. Additionally, this study presents a detailed overview of the CBIR framework and improvements achieved; including image preprocessing, feature extraction and indexing, system learning, benchmarking datasets, similarity matching, relevance feedback, performance evaluation, and visualization. Finally, promising research trends, challenges, and our insights are provided to inspire further research efforts.     

DEEP LEVEL DEFECTS IN SI-IMPLANTED LEC UNDOPED Si-GaAs

The residual electrically active defects in(4×10~(12)cm~(-2)(30KeV)+5×10~(12)cm~(-2)(130KeV))si-implanted LEC undoped si-GaAs activated by two-step rapid thermal annealing(RTA)LABELED AS 970℃(9S)+750℃(12S)have been investigated with deep level transient spec-troscopy(DLTS).Two electron traps ET_1(E_c-0.53eV,σ_n=2.3×10~(-16)cm~2)and ET_2(E_c-0.81eV,σ_n=9.7×10(-13)cm~2)are detected.Furthermore,the noticeable variations of trap's con-centration and energy level in the forbidden gap with the depth profile of defects induced by ion im-plantation and RTA process have also been observed.The[As_i·V_(As)·As_(Ga)]and[V_(As)·As_i·V_(Ga)·As_(Ga)]are proposed to be the possible atomic configurations of ET_1 and ET_2,respectively to explaintheir RTA behaviors.     

基于深对流云的夜间微光传感器定标技术

NPP(National Polar-orbiting Parternership)卫星于2011年10月28日发射升空,其上搭载的VIIRS(Visible Infrared Imaging Radiometer Suite)传感器包含DNB(Day Night Band)波段,该波段采用三增益设置,可在白天、晨昏甚至夜间实现地球观测.其中,在夜间月光条件下,主要采用高增益,定标实现较为困难.以此为目的,首次提出了基于深对流云的夜间微光传感器替代定标方法.将月亮辐射模型引入到SCIATRAN中,实现了全天候的辐射传输计算,并利用该改进的辐射传输模式,通过确定深对流云光学属性、地表特性及大气廓线等的参考值,对接收大气层顶反射辐射进行敏感性试验,分析得到主要影响因子为云光学厚度COT(Cloud Optical Thickness)与有效粒子半径Re(Effective Radius),且利用固定参考值得到最大模拟误差小于5%.为了检验该替代方法的可行性,通过改进的辐射传输模式,对2012年08月~2013年01月共6个月夜间DNB数据所确定的深对流云替代目标像素进行辐射传输模拟,结果表明基于深对流云的替代定标方法效果较好,基于日变化的辐射不确定度在±10%之内,可实现对DNB夜间高增益阶段的直接辐射定标.     

Molecular design of deep blue fluorescent emitters with 20% external quantum efficiency and narrow emission spectrum

A thermally activated delayed fluorescent (TADF) emitter, 2,7-bis(9,9-dimethylacridin-10(9H)-yl)-9,9-dimethyl-9H-thioxanthene 10,10-dioxide (DMTDAc), was developed as a deep blue TADF emitter using a rigid 9,9-dimethyl-9H-thioxanthene 10,10-dioxide (DMTD) acceptor and an acridine donor. The rigid DMTD acceptor narrowed emission spectrum of DMTDAc by interlocking two phenyl units of diphenylsulfone. A deep blue TADF device with an external quantum efficiency close to 20% with a deep blue color coordinate of (0.15,0.13) was provided using the DMTDAc TADF emitter.