It is difficult but important to get clear information from the low illumination images. In recent years the research of the low illumination image enhancement has become a hot topic in image processing and computer v...It is difficult but important to get clear information from the low illumination images. In recent years the research of the low illumination image enhancement has become a hot topic in image processing and computer vision. The Retinex algorithm is one of the most popular methods in the field and uniform illumination is necessary to enhance low illumination image quality by using this algorithm. However, for the different areas of an image with contrast brightness differences, the illumination image is not smooth and causes halo artifacts so that it cannot retain the detail information of the original images. To solve the problem, we generalize the multi-scale Retinex algorithm and propose a new enhancement method for the low illumination images based on the microarray camera. The proposed method can well make up for the deficiency of imbalanced illumination and significantly inhibit the halo artifacts as well. Experimental results show that the proposed method can get better image enhancement effect compared to the multi-scale Retinex algorithm of a single image enhancement. Advantages of the method also include that it can significantly inhibit the halo artifacts and thus retain the details of the original images, it can improve the brightness and contrast of the image as well. The newly developed method in this paper has application potential to the images captured by pad and cell phone in the low illumination environment.展开更多
The instabilities of indium–zinc oxide thin film transistors under bias and/or illumination stress are studied in this paper. Firstly, illumination experiments are performed, which indicates the variations of current...The instabilities of indium–zinc oxide thin film transistors under bias and/or illumination stress are studied in this paper. Firstly, illumination experiments are performed, which indicates the variations of current–voltage characteristics and electrical parameters(such as threshold voltage and sub-threshold swing) are dominated by the stress-induced ionized oxygen vacancies and acceptor-like states. The dependence of degradation on light wavelength is also investigated. More negative shift of threshold voltage and greater sub-threshold swing are observed with the decrease of light wavelength.Subsequently, a negative bias illumination stress experiment is carried out. The degradation of the device is aggravated due to the decrease of recombination effects between ionized oxygen vacancies and free carriers. Moreover, the contributions of ionized oxygen vacancies and acceptor-like states are separated by using the mid-gap method. In addition, ionized oxygen vacancies are partially recombined at room temperature and fully recombined at high temperature. Finally, low-frequency noise is measured before and after negative bias illumination stress. Experimental results show few variations of the oxide trapped charges are generated during stress, which is consistent with the proposed mechanism.展开更多
The sensing light source of the line scan camera cannot be fully exposed in a low light environment due to the extremely small number of photons and high noise,which leads to a reduction in image quality.A multi-scale...The sensing light source of the line scan camera cannot be fully exposed in a low light environment due to the extremely small number of photons and high noise,which leads to a reduction in image quality.A multi-scale fusion residual encoder-decoder(FRED)was proposed to solve the problem.By directly learning the end-to-end mapping between light and dark images,FRED can enhance the image’s brightness with the details and colors of the original image fully restored.A residual block(RB)was added to the network structure to increase feature diversity and speed up network training.Moreover,the addition of a dense context feature aggregation module(DCFAM)made up for the deficiency of spatial information in the deep network by aggregating the context’s global multi-scale features.The experimental results show that the FRED is superior to most other algorithms in visual effect and quantitative evaluation of peak signal-to-noise ratio(PSNR)and structural similarity index measure(SSIM).For the factor that FRED can restore the brightness of images while representing the edge and color of the image effectively,a satisfactory visual quality is obtained under the enhancement of low-light.展开更多
Structured illumination microscopy(SIM)is one of the most widely applied wide field super resolution imaging techniques with high temporal resolution and low phototoxicity.The spatial resolution of SIM is typically li...Structured illumination microscopy(SIM)is one of the most widely applied wide field super resolution imaging techniques with high temporal resolution and low phototoxicity.The spatial resolution of SIM is typically limited to two times of the diffraction limit and the depth of field is small.In this work,we propose and experimentally demonstrate a low cost,easy to implement,novel technique called speckle structured illumination endoscopy(SSIE)to enhance the resolution of a wide field endoscope with large depth of field.Here,speckle patterns are used to excite objects on the sample which is then followed by a blind-SIM algorithm for super resolution image reconstruction.Our approach is insensitive to the 3D morphology of the specimen,or the deformation of illuminations used.It greatly simplifies the experimental setup as there are no calibration protocols and no stringent control of illumination patterns nor focusing optics.We demonstrate that the SSIE can enhance the resolution 2–4.5 times that of a standard white light endoscopic(WLE)system.The SSIE presents a unique route to super resolution in endoscopic imaging at wide field of view and depth of field,which might be beneficial to the practice of clinical endoscopy.展开更多
随着深度学习方法的发展,理想环境下基于可见光的人脸识别精度和速度已经达到优秀的水平。但是在弱光等复杂环境下,由于缺少光源,可见光图像无法体现人脸细节,导致人脸识别效果下降甚至失效。为了解决这一问题,提出一种基于红外可见光...随着深度学习方法的发展,理想环境下基于可见光的人脸识别精度和速度已经达到优秀的水平。但是在弱光等复杂环境下,由于缺少光源,可见光图像无法体现人脸细节,导致人脸识别效果下降甚至失效。为了解决这一问题,提出一种基于红外可见光融合的复杂环境下人脸识别方法。首先,针对低照度环境提出联合CNN(Convolutional Neural Network)和Transformer的红外与可见光融合识别网络,并联CNN和视觉Transformer组成单模态特征融合模块,充分利用源图像的局部细节信息和全局上下文信息。同时,提出一种基于模态平均差异度的多模态特征融合策略,强化对源图像不同区域特征的差异化表达。其次,针对实际应用中融合识别网络模型大、速度慢的问题提出轻量化人脸识别网络Mobile Face Net-Coo和基于边云协同的自适应识别策略,通过图像质量选择识别模型,有效利用硬件资源。实验结果表明,弱光条件下,融合红外光与仅使用可见光图像相比,识别率提升了13.96个百分点。同时,将本方法应用实际项目中,结果表明:本方法在复杂环境下,能提高人脸识别的实时性和准确率。展开更多
基金Supported by National Science and Technology Major Project(2014ZX02502003The National Natural Science Foundation of China(61170327)
文摘It is difficult but important to get clear information from the low illumination images. In recent years the research of the low illumination image enhancement has become a hot topic in image processing and computer vision. The Retinex algorithm is one of the most popular methods in the field and uniform illumination is necessary to enhance low illumination image quality by using this algorithm. However, for the different areas of an image with contrast brightness differences, the illumination image is not smooth and causes halo artifacts so that it cannot retain the detail information of the original images. To solve the problem, we generalize the multi-scale Retinex algorithm and propose a new enhancement method for the low illumination images based on the microarray camera. The proposed method can well make up for the deficiency of imbalanced illumination and significantly inhibit the halo artifacts as well. Experimental results show that the proposed method can get better image enhancement effect compared to the multi-scale Retinex algorithm of a single image enhancement. Advantages of the method also include that it can significantly inhibit the halo artifacts and thus retain the details of the original images, it can improve the brightness and contrast of the image as well. The newly developed method in this paper has application potential to the images captured by pad and cell phone in the low illumination environment.
基金supported by the Opening Fund of Key Laboratory of Silicon Device Technology,Chinese Academy of Sciences(Grant No.KLSDTJJ2018-6)the National Natural Science Foundation of China(Grant No.61574048)+1 种基金the Science and Technology Research Project of Guangdong Province,China(Grant No.2015B090912002)the Pearl River S&T Nova Program of Guangzhou City,China(Grant No.201710010172)
文摘The instabilities of indium–zinc oxide thin film transistors under bias and/or illumination stress are studied in this paper. Firstly, illumination experiments are performed, which indicates the variations of current–voltage characteristics and electrical parameters(such as threshold voltage and sub-threshold swing) are dominated by the stress-induced ionized oxygen vacancies and acceptor-like states. The dependence of degradation on light wavelength is also investigated. More negative shift of threshold voltage and greater sub-threshold swing are observed with the decrease of light wavelength.Subsequently, a negative bias illumination stress experiment is carried out. The degradation of the device is aggravated due to the decrease of recombination effects between ionized oxygen vacancies and free carriers. Moreover, the contributions of ionized oxygen vacancies and acceptor-like states are separated by using the mid-gap method. In addition, ionized oxygen vacancies are partially recombined at room temperature and fully recombined at high temperature. Finally, low-frequency noise is measured before and after negative bias illumination stress. Experimental results show few variations of the oxide trapped charges are generated during stress, which is consistent with the proposed mechanism.
基金supported by the National Key Research and Development Project(2019YFC0121502)the Graduate Innovation Fund Project of Xi’an University of Posts and Telecommunications(CXJJLY2019062)the Key Laboratory of Network Data Analysis and Intelligent Processing of Shaanxi Province。
文摘The sensing light source of the line scan camera cannot be fully exposed in a low light environment due to the extremely small number of photons and high noise,which leads to a reduction in image quality.A multi-scale fusion residual encoder-decoder(FRED)was proposed to solve the problem.By directly learning the end-to-end mapping between light and dark images,FRED can enhance the image’s brightness with the details and colors of the original image fully restored.A residual block(RB)was added to the network structure to increase feature diversity and speed up network training.Moreover,the addition of a dense context feature aggregation module(DCFAM)made up for the deficiency of spatial information in the deep network by aggregating the context’s global multi-scale features.The experimental results show that the FRED is superior to most other algorithms in visual effect and quantitative evaluation of peak signal-to-noise ratio(PSNR)and structural similarity index measure(SSIM).For the factor that FRED can restore the brightness of images while representing the edge and color of the image effectively,a satisfactory visual quality is obtained under the enhancement of low-light.
基金partially supported by the Gordon and Betty Moore Foundation Grant No.5722
文摘Structured illumination microscopy(SIM)is one of the most widely applied wide field super resolution imaging techniques with high temporal resolution and low phototoxicity.The spatial resolution of SIM is typically limited to two times of the diffraction limit and the depth of field is small.In this work,we propose and experimentally demonstrate a low cost,easy to implement,novel technique called speckle structured illumination endoscopy(SSIE)to enhance the resolution of a wide field endoscope with large depth of field.Here,speckle patterns are used to excite objects on the sample which is then followed by a blind-SIM algorithm for super resolution image reconstruction.Our approach is insensitive to the 3D morphology of the specimen,or the deformation of illuminations used.It greatly simplifies the experimental setup as there are no calibration protocols and no stringent control of illumination patterns nor focusing optics.We demonstrate that the SSIE can enhance the resolution 2–4.5 times that of a standard white light endoscopic(WLE)system.The SSIE presents a unique route to super resolution in endoscopic imaging at wide field of view and depth of field,which might be beneficial to the practice of clinical endoscopy.
文摘随着深度学习方法的发展,理想环境下基于可见光的人脸识别精度和速度已经达到优秀的水平。但是在弱光等复杂环境下,由于缺少光源,可见光图像无法体现人脸细节,导致人脸识别效果下降甚至失效。为了解决这一问题,提出一种基于红外可见光融合的复杂环境下人脸识别方法。首先,针对低照度环境提出联合CNN(Convolutional Neural Network)和Transformer的红外与可见光融合识别网络,并联CNN和视觉Transformer组成单模态特征融合模块,充分利用源图像的局部细节信息和全局上下文信息。同时,提出一种基于模态平均差异度的多模态特征融合策略,强化对源图像不同区域特征的差异化表达。其次,针对实际应用中融合识别网络模型大、速度慢的问题提出轻量化人脸识别网络Mobile Face Net-Coo和基于边云协同的自适应识别策略,通过图像质量选择识别模型,有效利用硬件资源。实验结果表明,弱光条件下,融合红外光与仅使用可见光图像相比,识别率提升了13.96个百分点。同时,将本方法应用实际项目中,结果表明:本方法在复杂环境下,能提高人脸识别的实时性和准确率。