In this study,the mechanism and characteristics of the responseαparticles and the damage caused by them in CMOS active pixel(APS)sensors were investigated.A detection and compensation algorithm for dead pixels caused...In this study,the mechanism and characteristics of the responseαparticles and the damage caused by them in CMOS active pixel(APS)sensors were investigated.A detection and compensation algorithm for dead pixels caused byαparticle ionizing radiation was proposed,and the effects of dead-pixel compensation algorithms were compared and analyzed under different parameter conditions.The experimental results show thatαparticle response signal has highest accuracy at 9 dB gain,with an obvious“target-ring”distribution.With increasing cumulative dose,the CMOS APS pedestal tends to saturation while dead pixels continue increasing.Though some pixel damage recovers through natural annealing,the dead-to-noise ratio increases with irradiation time,reaching 32.54%after 72 h.A hierarchical clustering dead-pixel detection method is proposed,categorizing pixels into two types:those within and outside the response event.A classification compensation strategy combining mean and majority filtering is proposed.This compensation algorithm can address dead-pixel interference without affectingαparticle radiation response data.When iterated multiple times and with integration time exceeding 6.31 ms,the number of dead pixels can be effectively reduced.展开更多
目的:探讨玫瑰痤疮患者采用低能量Pixel调Q像束激光与米诺环素联合治疗后红斑情况及面部功能改善情况。方法:研究对象选自于2020年3月-2023年7月在上海交通大学医学院苏州九龙医院接受治疗的玫瑰痤疮患者110例,并分为对照组和观察组,每...目的:探讨玫瑰痤疮患者采用低能量Pixel调Q像束激光与米诺环素联合治疗后红斑情况及面部功能改善情况。方法:研究对象选自于2020年3月-2023年7月在上海交通大学医学院苏州九龙医院接受治疗的玫瑰痤疮患者110例,并分为对照组和观察组,每组例数均为55例,分组方法为随机数字表法。对照组给予盐酸米诺环素片,在对照组患者治疗的基础上,采用低能量Pixel调Q像束激光对观察组患者进行治疗。两组治疗时间均为6周。比较两组临床疗效(治疗6周后),玫瑰痤疮情况、整体病情、p38丝裂原活化蛋白激酶(p38 mitogen activated protein kinase,p38MAPK)通路蛋白、生活质量、红斑情况、面部功能、炎性因子(治疗前、治疗6周后),不良反应(治疗期间)。结果:治疗6周后,观察组总有效率高于对照组(76.36%vs.54.55%,P<0.05)。与治疗前比较,两组治疗6周后的玫瑰痤疮医师全球评分(Patient's global assessment,PGA)、整体病情评估(Investigator global assessment,IGA)、皮肤病生活质量量表(Dermatology life quality index,DLQI)评分、红斑评估量表(Clinician′s erythema assessment,CEA)评分、皮损区红斑指数(Erythema index,EI)、经皮水分丢失(Trans epidermal water loss,TEWL)、皮肤酸碱度(Pondus Hydrogenii,pH)值、炎性因子及p38MAPK通路蛋白相关因子水平均降低,且相比于对照组,观察组更低;两组角质层含水量均升高,且相比于对照组,观察组更高(P<0.05)。观察组和对照组治疗期间的不良反应发生率比较差异无统计学意义(P>0.05)。结论:玫瑰痤疮患者经低能量Pixel调Q像束激光联合米诺环素治疗后,其相关临床症状可得到有效缓解,炎症反应得以减轻,并可调节p38MAPK信号通路蛋白的表达,进一步可促进患者红斑情况及面部功能的改善,进而促使患者生活质量及临床疗效得以提高,且具有良好的安全性。展开更多
Brazil’s deforestation monitoring integrates accuracy and current monitoring for land use and land cover applications.Regular monitoring of deforestation and non-deforestation requires Sentinel-2 multispectral satell...Brazil’s deforestation monitoring integrates accuracy and current monitoring for land use and land cover applications.Regular monitoring of deforestation and non-deforestation requires Sentinel-2 multispectral satellite images of several bands at various frequencies,the mix of high-and low-resolution images that make object classification difficult because of the mixed pixel problem.Accuracy is impacted by the mixed pixel problem,which occurs when pixels belong to different classes and makes detection challenging.To identify mixed pixels,Band Math is used to merge numerous bands to generate a new band NDVI.Thresholding is used to analyze the edges of deforested and non-deforested areas.Segmentation is then used to analyze the pixels which helps to identify the number of mixed pixels to compute the deforested and non-deforested areas.Segmented image pixels are used to categorize the deforestation of the Brazilian Amazon Forest between 2019 and 2023.Verify how many pixels are mixed to improve accuracy and identify mixed pixel issues;compare the mixed and pure pixels of fuzzy clustering with the subtracted morphological image pixels.With the help of segmentation and clustering researchers effectively validate mixed pixels in a specific area.The proposed methodology is easy to analyze and helpful for an appropriate calculation of deforested and non-deforested areas.展开更多
Maritime target recognition and image perception enhancement are gradually being promoted and applied in ocean engineering. This paper proposes the attentional multi-pixel fusion(AMF) algorithm for the intelligent nav...Maritime target recognition and image perception enhancement are gradually being promoted and applied in ocean engineering. This paper proposes the attentional multi-pixel fusion(AMF) algorithm for the intelligent navigation of unmanned surface vessels(USVs). The algorithm preprocesses the image pixel matrix in blocks, computes the mapping between regional and full-pixel matrices, and adaptively equalizes the mapping weights via a Gaussian-fuzzy matrix.This approach guarantees the preservation of the target contour and texture information. Compared with five classic enhancement algorithms, the AMF algorithm improves the peak signal-to-noise ratio(PSNR) and structural similarity index(SSIM). Experimental validation via YOLOv8 for maritime target detection demonstrates 2.1% and 2.4%improvements in the evaluation indices over training on 4000 original images, with shorter training times and lower confusion rates. In maritime target ranging, the AMF algorithm, coupled with the ISR method, exhibits the lowest improved stereo ranging mean absolute error and standard deviation values and higher similarity between the regional and full-pixel matrices. In summary, the AMF algorithm excels in target detection and ranging, offering promising applications in ocean engineering, such as marine resource exploitation, path planning, and intelligent collaboration among unmanned vessels.展开更多
In hyperspectral image classification(HSIC),accurately extracting spatial and spectral information from hyperspectral images(HSI)is crucial for achieving precise classification.However,due to low spatial resolution an...In hyperspectral image classification(HSIC),accurately extracting spatial and spectral information from hyperspectral images(HSI)is crucial for achieving precise classification.However,due to low spatial resolution and complex category boundary,mixed pixels containing features from multiple classes are inevitable in HSIs.Additionally,the spectral similarity among different classes challenge for extracting distinctive spectral features essential for HSIC.To address the impact of mixed pixels and spectral similarity for HSIC,we propose a central-pixel guiding sub-pixel and sub-channel convolution network(CP-SPSC)to extract more precise spatial and spectral features.Firstly,we designed spatial attention(CP-SPA)and spectral attention(CP-SPE)informed by the central pixel to effectively reduce spectral interference of irrelevant categories in the same patch.Furthermore,we use CP-SPA to guide 2D sub-pixel convolution(SPConv2d)to capture spatial features finer than the pixel level.Meanwhile,CP-SPE is also utilized to guide 1D sub-channel con-volution(SCConv1d)in selecting more precise spectral channels.For fusing spatial and spectral information at the feature-level,the spectral feature extension transformation module(SFET)adopts mirror-padding and snake permutation to transform 1D spectral information of the center pixel into 2D spectral features.Experiments on three popular datasets demonstrate that ours out-performs several state-of-the-art methods in accuracy.展开更多
The types and structures of inorganic pores are key factors in evaluations of the reservoir space and distribution characteristics of shale oil and gas.However,quantitative identification methods for pores of differen...The types and structures of inorganic pores are key factors in evaluations of the reservoir space and distribution characteristics of shale oil and gas.However,quantitative identification methods for pores of different inorganic components have not yet been fully developed.For this reason,a quantitative characterization method of inorganic pores using pixel information was proposed in this study.A machine learning algorithm was used to assist the field emission scanning electron microscopy(FE-SEM)image processing of shale to realize the accurate identification and quantitative characterization of inorganic pores on the surface of high-precision images of shale with a small view.Moreover,large-view image splicing technology,combined with quantitative evaluation of minerals by scanning electron microscopy(QEMSCAN)image joint characterization technology,was used to accurately analyze the distribution characteristics of inorganic pores under different mineral components.The quantitative methods of pore characteristics of different inorganic components under the pixel information of shale were studied.The results showed that(1)the Waikato Environment for Knowledge Analysis(WEKA)machine learning model can effectively identify and extract shale mineral components and inorganic pore distribution,and the large-view FE-SEM images are representative of samples at the 200μm×200μm view scale,meeting statistical requirements and eliminating the influence of heterogeneity;(2)the pores developed by different mineral components of shale had obvious differences,indicating that the development of inorganic pores is highly correlated with the properties of shale minerals themselves;and(3)the pore-forming ability of different mineral components is calculated by the quantitative method of single component pore-forming coefficient.Chlorite showed the highest pore-forming ability,followed by(in descending order)illite,pyrite,calcite,dolomite,albite,orthoclase,quartz,and apatite.This study contributes to advancing our understanding of inorganic pore characteristics in shale.展开更多
In the field of imaging,the image resolution is required to be higher.There is always a contradiction between the sensitivity and resolution of the seeker in the infrared guidance system.This work uses the rosette sca...In the field of imaging,the image resolution is required to be higher.There is always a contradiction between the sensitivity and resolution of the seeker in the infrared guidance system.This work uses the rosette scanning mode for physical compression imaging in order to improve the resolution of the image as much as possible under the high-sensitivity infrared rosette point scanning mode and complete the missing information that is not scanned.It is effective to use optical lens instead of traditional optical reflection system,which can reduce the loss in optical path transmission.At the same time,deep learning neural network is used for control.An infrared single pixel imaging system that integrates sparse algorithm and recovery algorithm through the improved generative adversarial networks is trained.The experiment on the infrared aerial target dataset shows that when the input is sparse image after rose sampling,the system finally can realize the single pixel recovery imaging of the infrared image,which improves the resolution of the image while ensuring high sensitivity.展开更多
Metamaterials have exotic physical properties that rely on the construction of their underlying architecture.However,the physical properties of conventional mechanical metamaterials are permanently programmed into the...Metamaterials have exotic physical properties that rely on the construction of their underlying architecture.However,the physical properties of conventional mechanical metamaterials are permanently programmed into their periodic interconnect configurations,resulting in their lack of modularity,scalable fabrication,and programmability.Mechanical metamaterials typically exhibit a single extraordinary mechanical property or multiple extraordinary properties coupled together,making it difficult to realize multiple independent extraordinary mechanical properties.Here,the pixel mechanics metamaterials(PMMs)with multifunctional and reprogrammable properties are developed by arraying uncoupled constrained individual modular mechanics pixels(MPs).The MPs enable controlled conversion between two extraordinary mechanical properties(multistability and compression-torsion coupling deformation).Each MP exhibits 32 independent and reversible room temperature programming configurations.In addition,the programmability of metamaterials is further enhanced by shape memory polymer(SMP)and 4D printing,greatly enriching the design freedom.For the PMM consisting of m×n MPs,it has 32(m×n)independent room temperature programming configurations.The application prospects of metamaterials in the vibration isolation device and energy absorption device with programmable performance have been demonstrated.The vibration isolation frequencies of the MP before and after programming were[0 Hz-5.86 Hz],[0 Hz-13.67 Hz and 306.64 Hz-365.23 Hz].The total energy absorption of the developed PMM can be adjusted controllably in the range of 1.01 J-3.91 J.Six standard digital logic gates that do not require sustained external force are designed by controlling the closure between the modules.This design paradigm will facilitate the further development of multifunctional and reprogrammable metamaterials.展开更多
Previous research utilizing Cartoon Generative Adversarial Network(CartoonGAN)has encountered limitations in managing intricate outlines and accurately representing lighting effects,particularly in complex scenes requ...Previous research utilizing Cartoon Generative Adversarial Network(CartoonGAN)has encountered limitations in managing intricate outlines and accurately representing lighting effects,particularly in complex scenes requiring detailed shading and contrast.This paper presents a novel Enhanced Pixel Integration(EPI)technique designed to improve the visual quality of images generated by CartoonGAN.Rather than modifying the core model,the EPI approach employs post-processing adjustments that enhance images without significant computational overhead.In this method,images produced by CartoonGAN are converted from Red-Green-Blue(RGB)to Hue-Saturation-Value(HSV)format,allowing for precise adjustments in hue,saturation,and brightness,thereby improving color fidelity.Specific correction values are applied to fine-tune colors,ensuring they closely match the original input while maintaining the characteristic,stylized effect of CartoonGAN.The corrected images are blended with the originals to retain aesthetic appeal and visual distinctiveness,resulting in improved color accuracy and overall coherence.Experimental results demonstrate that EPI significantly increases similarity to original input images compared to the standard CartoonGAN model,achieving a 40.14%enhancement in visual similarity in Learned Perceptual Image Patch Similarity(LPIPS),a 30.21%improvement in structural consistency in Structural Similarity Index Measure(SSIM),and an 11.81%reduction in pixel-level error in Mean Squared Error(MSE).By addressing limitations present in the traditional CartoonGAN pipeline,EPI offers practical enhancements for creative applications,particularly within media and design fields where visual fidelity and artistic style preservation are critical.These improvements align with the goals of Fog and Edge Computing,which also seek to enhance processing efficiency and application performance in sensitive industries such as healthcare,logistics,and education.This research not only resolves key deficiencies in existing CartoonGAN models but also expands its potential applications in image-based content creation,bridging gaps between technical constraints and creative demands.Future studies may explore the adaptability of EPI across various datasets and artistic styles,potentially broadening its impact on visual transformation tasks.展开更多
In recent years,deep learning has been introduced into the field of Single-pixel imaging(SPI),garnering significant attention.However,conventional networks still exhibit limitations in preserving image details.To addr...In recent years,deep learning has been introduced into the field of Single-pixel imaging(SPI),garnering significant attention.However,conventional networks still exhibit limitations in preserving image details.To address this issue,we integrate Large Kernel Convolution(LKconv)into the U-Net framework,proposing an enhanced network structure named U-LKconv network,which significantly enhances the capability to recover image details even under low sampling conditions.展开更多
A three-transistor active pixel sensor and its double sampling readout circuit implemented by a switch capacitor amplifier are designed. The circuit is embedded in a 64 × 64 pixel array CMOS image sensor and succ...A three-transistor active pixel sensor and its double sampling readout circuit implemented by a switch capacitor amplifier are designed. The circuit is embedded in a 64 × 64 pixel array CMOS image sensor and success-fully taped out with a Chartered 0.35μm process. The pixel pitch is 8μm × 8μm with a fill factor of 57%, the photo-sensitivity is 0.8V/（lux · s） ,and the dynamic range is 50dB. Theoretical analysis and test results indicate that as the process is scaled down, a smaller pixel pitch reduces the sensitivity. A deep junction n-well/p-substrate photodiode with a reasonable fill factor and high sensitivity are more appropriate for submicron processes.展开更多
We present a novel scheme for embedding secret data into a binary image without introducing noticeable artifacts. Unlike some block-based methods, the proposed scheme encodes the secret bits directly into boundary pix...We present a novel scheme for embedding secret data into a binary image without introducing noticeable artifacts. Unlike some block-based methods, the proposed scheme encodes the secret bits directly into boundary pixels by checking each pixel of the cover image in a pseudo-random order for embedding eligibility. A set of rules ensures correct identification of data-carrying pixels in blind extraction. The proposed scheme does not generate isolated dots, and can incorporate various coding methods such as matrix encoding to further improve the embedding performance. It is shown that up to one fourth of the boundary pixels may be used to carry secret data. Experimental results indicate that the method can achieve good visual quality with fairly large data capacity.展开更多
基金supported by the National Natural Science Foundation of China(No.11905102)Hunan Provincial Postgraduate Research and Innovation Project(No.QL20230234)。
文摘In this study,the mechanism and characteristics of the responseαparticles and the damage caused by them in CMOS active pixel(APS)sensors were investigated.A detection and compensation algorithm for dead pixels caused byαparticle ionizing radiation was proposed,and the effects of dead-pixel compensation algorithms were compared and analyzed under different parameter conditions.The experimental results show thatαparticle response signal has highest accuracy at 9 dB gain,with an obvious“target-ring”distribution.With increasing cumulative dose,the CMOS APS pedestal tends to saturation while dead pixels continue increasing.Though some pixel damage recovers through natural annealing,the dead-to-noise ratio increases with irradiation time,reaching 32.54%after 72 h.A hierarchical clustering dead-pixel detection method is proposed,categorizing pixels into two types:those within and outside the response event.A classification compensation strategy combining mean and majority filtering is proposed.This compensation algorithm can address dead-pixel interference without affectingαparticle radiation response data.When iterated multiple times and with integration time exceeding 6.31 ms,the number of dead pixels can be effectively reduced.
文摘目的:探讨玫瑰痤疮患者采用低能量Pixel调Q像束激光与米诺环素联合治疗后红斑情况及面部功能改善情况。方法:研究对象选自于2020年3月-2023年7月在上海交通大学医学院苏州九龙医院接受治疗的玫瑰痤疮患者110例,并分为对照组和观察组,每组例数均为55例,分组方法为随机数字表法。对照组给予盐酸米诺环素片,在对照组患者治疗的基础上,采用低能量Pixel调Q像束激光对观察组患者进行治疗。两组治疗时间均为6周。比较两组临床疗效(治疗6周后),玫瑰痤疮情况、整体病情、p38丝裂原活化蛋白激酶(p38 mitogen activated protein kinase,p38MAPK)通路蛋白、生活质量、红斑情况、面部功能、炎性因子(治疗前、治疗6周后),不良反应(治疗期间)。结果:治疗6周后,观察组总有效率高于对照组(76.36%vs.54.55%,P<0.05)。与治疗前比较,两组治疗6周后的玫瑰痤疮医师全球评分(Patient's global assessment,PGA)、整体病情评估(Investigator global assessment,IGA)、皮肤病生活质量量表(Dermatology life quality index,DLQI)评分、红斑评估量表(Clinician′s erythema assessment,CEA)评分、皮损区红斑指数(Erythema index,EI)、经皮水分丢失(Trans epidermal water loss,TEWL)、皮肤酸碱度(Pondus Hydrogenii,pH)值、炎性因子及p38MAPK通路蛋白相关因子水平均降低,且相比于对照组,观察组更低;两组角质层含水量均升高,且相比于对照组,观察组更高(P<0.05)。观察组和对照组治疗期间的不良反应发生率比较差异无统计学意义(P>0.05)。结论:玫瑰痤疮患者经低能量Pixel调Q像束激光联合米诺环素治疗后,其相关临床症状可得到有效缓解,炎症反应得以减轻,并可调节p38MAPK信号通路蛋白的表达,进一步可促进患者红斑情况及面部功能的改善,进而促使患者生活质量及临床疗效得以提高,且具有良好的安全性。
文摘Brazil’s deforestation monitoring integrates accuracy and current monitoring for land use and land cover applications.Regular monitoring of deforestation and non-deforestation requires Sentinel-2 multispectral satellite images of several bands at various frequencies,the mix of high-and low-resolution images that make object classification difficult because of the mixed pixel problem.Accuracy is impacted by the mixed pixel problem,which occurs when pixels belong to different classes and makes detection challenging.To identify mixed pixels,Band Math is used to merge numerous bands to generate a new band NDVI.Thresholding is used to analyze the edges of deforested and non-deforested areas.Segmentation is then used to analyze the pixels which helps to identify the number of mixed pixels to compute the deforested and non-deforested areas.Segmented image pixels are used to categorize the deforestation of the Brazilian Amazon Forest between 2019 and 2023.Verify how many pixels are mixed to improve accuracy and identify mixed pixel issues;compare the mixed and pure pixels of fuzzy clustering with the subtracted morphological image pixels.With the help of segmentation and clustering researchers effectively validate mixed pixels in a specific area.The proposed methodology is easy to analyze and helpful for an appropriate calculation of deforested and non-deforested areas.
基金financially supported by the Foundation of Shanxi Key Laboratory of Machine Vision and Virtual Reality (Grant No.447-110103)the Science and Technology Innovation Plan of Shanghai Science and Technology Commission (Grant No. 22dz1204000)。
文摘Maritime target recognition and image perception enhancement are gradually being promoted and applied in ocean engineering. This paper proposes the attentional multi-pixel fusion(AMF) algorithm for the intelligent navigation of unmanned surface vessels(USVs). The algorithm preprocesses the image pixel matrix in blocks, computes the mapping between regional and full-pixel matrices, and adaptively equalizes the mapping weights via a Gaussian-fuzzy matrix.This approach guarantees the preservation of the target contour and texture information. Compared with five classic enhancement algorithms, the AMF algorithm improves the peak signal-to-noise ratio(PSNR) and structural similarity index(SSIM). Experimental validation via YOLOv8 for maritime target detection demonstrates 2.1% and 2.4%improvements in the evaluation indices over training on 4000 original images, with shorter training times and lower confusion rates. In maritime target ranging, the AMF algorithm, coupled with the ISR method, exhibits the lowest improved stereo ranging mean absolute error and standard deviation values and higher similarity between the regional and full-pixel matrices. In summary, the AMF algorithm excels in target detection and ranging, offering promising applications in ocean engineering, such as marine resource exploitation, path planning, and intelligent collaboration among unmanned vessels.
基金supported by the National Natural Science Foundation of China(No.62071323).
文摘In hyperspectral image classification(HSIC),accurately extracting spatial and spectral information from hyperspectral images(HSI)is crucial for achieving precise classification.However,due to low spatial resolution and complex category boundary,mixed pixels containing features from multiple classes are inevitable in HSIs.Additionally,the spectral similarity among different classes challenge for extracting distinctive spectral features essential for HSIC.To address the impact of mixed pixels and spectral similarity for HSIC,we propose a central-pixel guiding sub-pixel and sub-channel convolution network(CP-SPSC)to extract more precise spatial and spectral features.Firstly,we designed spatial attention(CP-SPA)and spectral attention(CP-SPE)informed by the central pixel to effectively reduce spectral interference of irrelevant categories in the same patch.Furthermore,we use CP-SPA to guide 2D sub-pixel convolution(SPConv2d)to capture spatial features finer than the pixel level.Meanwhile,CP-SPE is also utilized to guide 1D sub-channel con-volution(SCConv1d)in selecting more precise spectral channels.For fusing spatial and spectral information at the feature-level,the spectral feature extension transformation module(SFET)adopts mirror-padding and snake permutation to transform 1D spectral information of the center pixel into 2D spectral features.Experiments on three popular datasets demonstrate that ours out-performs several state-of-the-art methods in accuracy.
基金supported by the National Natural Science Foundation of China(42372144)the Natural Science Foundation of Xinjiang Uygur Autonomous Region(2024D01E09)the Strategic Cooperation Technology Projects of CNPC and CUPB(ZLZX2020-01-05).
文摘The types and structures of inorganic pores are key factors in evaluations of the reservoir space and distribution characteristics of shale oil and gas.However,quantitative identification methods for pores of different inorganic components have not yet been fully developed.For this reason,a quantitative characterization method of inorganic pores using pixel information was proposed in this study.A machine learning algorithm was used to assist the field emission scanning electron microscopy(FE-SEM)image processing of shale to realize the accurate identification and quantitative characterization of inorganic pores on the surface of high-precision images of shale with a small view.Moreover,large-view image splicing technology,combined with quantitative evaluation of minerals by scanning electron microscopy(QEMSCAN)image joint characterization technology,was used to accurately analyze the distribution characteristics of inorganic pores under different mineral components.The quantitative methods of pore characteristics of different inorganic components under the pixel information of shale were studied.The results showed that(1)the Waikato Environment for Knowledge Analysis(WEKA)machine learning model can effectively identify and extract shale mineral components and inorganic pore distribution,and the large-view FE-SEM images are representative of samples at the 200μm×200μm view scale,meeting statistical requirements and eliminating the influence of heterogeneity;(2)the pores developed by different mineral components of shale had obvious differences,indicating that the development of inorganic pores is highly correlated with the properties of shale minerals themselves;and(3)the pore-forming ability of different mineral components is calculated by the quantitative method of single component pore-forming coefficient.Chlorite showed the highest pore-forming ability,followed by(in descending order)illite,pyrite,calcite,dolomite,albite,orthoclase,quartz,and apatite.This study contributes to advancing our understanding of inorganic pore characteristics in shale.
基金the Fundamental Research Funds for the Central Universities(No.3072022CF0802)。
文摘In the field of imaging,the image resolution is required to be higher.There is always a contradiction between the sensitivity and resolution of the seeker in the infrared guidance system.This work uses the rosette scanning mode for physical compression imaging in order to improve the resolution of the image as much as possible under the high-sensitivity infrared rosette point scanning mode and complete the missing information that is not scanned.It is effective to use optical lens instead of traditional optical reflection system,which can reduce the loss in optical path transmission.At the same time,deep learning neural network is used for control.An infrared single pixel imaging system that integrates sparse algorithm and recovery algorithm through the improved generative adversarial networks is trained.The experiment on the infrared aerial target dataset shows that when the input is sparse image after rose sampling,the system finally can realize the single pixel recovery imaging of the infrared image,which improves the resolution of the image while ensuring high sensitivity.
基金the financial support provided by the National Key R&D Program of China(2022YFB3805700)the National Natural Science Foundation of China(Grant Nos.12072094 and 12172106)+2 种基金the China Postdoctoral Science Foundation(Grant No.2023M730869)the Heilongjiang Natural Science Foundation Joint Guidance Project(Grant No.LH2023A004)the Postdoctoral Fellowship Program of CPSF(Grant No.GZB20230959)。
文摘Metamaterials have exotic physical properties that rely on the construction of their underlying architecture.However,the physical properties of conventional mechanical metamaterials are permanently programmed into their periodic interconnect configurations,resulting in their lack of modularity,scalable fabrication,and programmability.Mechanical metamaterials typically exhibit a single extraordinary mechanical property or multiple extraordinary properties coupled together,making it difficult to realize multiple independent extraordinary mechanical properties.Here,the pixel mechanics metamaterials(PMMs)with multifunctional and reprogrammable properties are developed by arraying uncoupled constrained individual modular mechanics pixels(MPs).The MPs enable controlled conversion between two extraordinary mechanical properties(multistability and compression-torsion coupling deformation).Each MP exhibits 32 independent and reversible room temperature programming configurations.In addition,the programmability of metamaterials is further enhanced by shape memory polymer(SMP)and 4D printing,greatly enriching the design freedom.For the PMM consisting of m×n MPs,it has 32(m×n)independent room temperature programming configurations.The application prospects of metamaterials in the vibration isolation device and energy absorption device with programmable performance have been demonstrated.The vibration isolation frequencies of the MP before and after programming were[0 Hz-5.86 Hz],[0 Hz-13.67 Hz and 306.64 Hz-365.23 Hz].The total energy absorption of the developed PMM can be adjusted controllably in the range of 1.01 J-3.91 J.Six standard digital logic gates that do not require sustained external force are designed by controlling the closure between the modules.This design paradigm will facilitate the further development of multifunctional and reprogrammable metamaterials.
基金supported by the National Research Foundation of Korea(NRF)under Grant RS-2022-NR-069955(2022R1A2C1092178).
文摘Previous research utilizing Cartoon Generative Adversarial Network(CartoonGAN)has encountered limitations in managing intricate outlines and accurately representing lighting effects,particularly in complex scenes requiring detailed shading and contrast.This paper presents a novel Enhanced Pixel Integration(EPI)technique designed to improve the visual quality of images generated by CartoonGAN.Rather than modifying the core model,the EPI approach employs post-processing adjustments that enhance images without significant computational overhead.In this method,images produced by CartoonGAN are converted from Red-Green-Blue(RGB)to Hue-Saturation-Value(HSV)format,allowing for precise adjustments in hue,saturation,and brightness,thereby improving color fidelity.Specific correction values are applied to fine-tune colors,ensuring they closely match the original input while maintaining the characteristic,stylized effect of CartoonGAN.The corrected images are blended with the originals to retain aesthetic appeal and visual distinctiveness,resulting in improved color accuracy and overall coherence.Experimental results demonstrate that EPI significantly increases similarity to original input images compared to the standard CartoonGAN model,achieving a 40.14%enhancement in visual similarity in Learned Perceptual Image Patch Similarity(LPIPS),a 30.21%improvement in structural consistency in Structural Similarity Index Measure(SSIM),and an 11.81%reduction in pixel-level error in Mean Squared Error(MSE).By addressing limitations present in the traditional CartoonGAN pipeline,EPI offers practical enhancements for creative applications,particularly within media and design fields where visual fidelity and artistic style preservation are critical.These improvements align with the goals of Fog and Edge Computing,which also seek to enhance processing efficiency and application performance in sensitive industries such as healthcare,logistics,and education.This research not only resolves key deficiencies in existing CartoonGAN models but also expands its potential applications in image-based content creation,bridging gaps between technical constraints and creative demands.Future studies may explore the adaptability of EPI across various datasets and artistic styles,potentially broadening its impact on visual transformation tasks.
文摘In recent years,deep learning has been introduced into the field of Single-pixel imaging(SPI),garnering significant attention.However,conventional networks still exhibit limitations in preserving image details.To address this issue,we integrate Large Kernel Convolution(LKconv)into the U-Net framework,proposing an enhanced network structure named U-LKconv network,which significantly enhances the capability to recover image details even under low sampling conditions.
文摘A three-transistor active pixel sensor and its double sampling readout circuit implemented by a switch capacitor amplifier are designed. The circuit is embedded in a 64 × 64 pixel array CMOS image sensor and success-fully taped out with a Chartered 0.35μm process. The pixel pitch is 8μm × 8μm with a fill factor of 57%, the photo-sensitivity is 0.8V/（lux · s） ,and the dynamic range is 50dB. Theoretical analysis and test results indicate that as the process is scaled down, a smaller pixel pitch reduces the sensitivity. A deep junction n-well/p-substrate photodiode with a reasonable fill factor and high sensitivity are more appropriate for submicron processes.
基金Project supported by the National Natural Science Foundation of China （Grant Nos.60372090, 60502039）, and the Key Project of Shanghai Municipality for Basic Research （Grant No.04JC14037）
文摘We present a novel scheme for embedding secret data into a binary image without introducing noticeable artifacts. Unlike some block-based methods, the proposed scheme encodes the secret bits directly into boundary pixels by checking each pixel of the cover image in a pseudo-random order for embedding eligibility. A set of rules ensures correct identification of data-carrying pixels in blind extraction. The proposed scheme does not generate isolated dots, and can incorporate various coding methods such as matrix encoding to further improve the embedding performance. It is shown that up to one fourth of the boundary pixels may be used to carry secret data. Experimental results indicate that the method can achieve good visual quality with fairly large data capacity.
