最新刊期
卷 41 ， 期 9 ， 2026
- 摘要：To address the challenge of characterizing the thickness of polyimide （PI） alignment films in the peripheral display area of TFT-LCDs， a film thickness distribution characterization method based on time-of-flight secondary ion mass spectrometry （TOF-SIMS） depth analysis is proposed. An argon gas cluster ion beam（GCIB） was employed to sputter the inkjet-printed PI films at multiple micro-regions. The substrate-derived InO2 was selected as the characteristic ion， and the sputtering time corresponding to the 50% midpoint of its normalized signal intensity was defined as the breakthrough sputtering time t. The sputtering times across 17 micro-regions from the film edge to the central area were measured. Independent step profiler measurements of sputter crater depths and original film thicknesses were used for validation， together with a quantitative statistical analysis of the correlation between sputtering time and film thickness. The method was applied to four PI film samples prepared under different pre-baking conditions. The sputtering time obtained by the TOF-SIMS depth profiling method is highly consistent with the spatial distribution trend of the original film thickness measured by stylus profilometry（overall Spearman correlation coefficients of 0.571~0.963）. Samples A and C， subjected to quiescent time， exhibited significant thickness anomalies in the edge region， with the maximum sputter times reaching 378 s and 487 s respectively， far exceeding the average of approximately 180 s. Samples B and D， without quiescent time， maintained uniform sputter times （~180 s） across the entire measured area. This method provides an effective analytical tool for the characterization of film uniformity across multiple microscopic regions， offering technical support for inkjet-printed PI alignment film process optimization and narrow-bezel panel manufacturing.关键词：time-of-flight secondary ion mass spectrometry;depth profile;polyimide;liquid crystal alignment film;film thickness characterization;inkjet printing4|1|0更新时间：2026-10-16
Material Physics
- 摘要：Various achromatic planar optical elements， such as q-plates and polarization gratings， can be designed and fabricated based on the geometric phase of liquid crystals. The function of these optical elements heavily relies on the precise alignment of liquid crystal molecules. While digital micromirror device（DMD） exposure has been widely used for the photoalignment of liquid crystal geometric phase elements， its dimensional resolution is inherently limited by the micromirror pitch and the magnification of the projection lens. To address these limitations， this paper proposes a polarized laser direct writing （PLDW） system which integrates a scanning galvanometer with a displacement stage. The developed PLDW system achieves a high dimension resolution of 3.4 μm and demonstrates effective photoalignment capabilities. Utilizing this system， Pancharatnam-Berry（PB） gratings， lens and microlens arrays were fabricated. The fabricated PB grating exhibits a fine period of 27 μm and achieves a relative diffraction efficiency exceeding 85% under circularly polarized illumination. Furthermore， the PB lens demonstrated good focusing performance， and the microlens array successfully generated a clear image array of the mask pattern at a magnification of 1∶10. The proposed PLDW technique offers a high-precision and cost-effective solution for the photoalignment of liquid crystal geometric phase planar optical elements.关键词：laser direct writing;photoalignment;geometric phase;Liquid crystal optical elements3|0|0更新时间：2026-10-16
- 摘要：Conventional switchable liquid crystal（LC） gratings generally suffer from polarization sensitivity， poor continuous electrical tunability of the grating period， and slow response， which cannot meet the requirements for multi-dimensional and high precision light manipulation in complex optical systems. To address these drawbacks， this work adopts a flexoelectric LC mixture with a broad operating temperature range （0~65 ℃）. Combined with patterned photoalignment， we fabricate a dual-mode switchable LC grating capable of continuous tuning of grating period， diffraction angle， and diffraction efficiency. The grating contains two periodic structures with markedly different feature sizes and can generate two distinct diffraction signals at different propagation distances. At zero applied voltage， transverse one-dimensional diffraction with small diffraction angles originating from the photo aligned pattern is observed. When the applied voltage exceeds 14 V， cross-stripe diffraction of large angles originating from the flexoelectric effect is activated. Within the voltage range of 14 V to 26 V， the grating period can be continuously tuned from 6.65 µm to 1.87 µm. The tuning for full range of period takes approximately 100 s， and the tuning speed is nearly 12 times higher than that of conventional light-driven cholesteric LC gratings. Meanwhile， the device exhibits weak polarization dependence and supports simultaneous continuous electrical modulation of grating period， diffraction angle， and diffraction efficiency. It features a simple fabrication procedure without elaborate patterned electrode structures. It holds promising application prospects in dynamic optical scenarios such as spectral scanning， adaptive beam shaping， and augmented reality displays， and provides a new technical route for the design of tunable LC optoelectronic devices.关键词：flexoelectric effect;broad temperature range;continuous period tuning;mode-switching grating;patterned photoalignment2|0|0更新时间：2026-10-16
Device Physics and Device Preparation
- 摘要：Flexoelectric effect is a critical factor causing image flicker in fringe-field switching liquid crystal displays（FFS-LCDs） under low-frequency driving， yet accurate measurement of flexoelectric coefficients （
- 摘要：Organic light-emitting diode（OLED） displays are subject to uneven brightness decay and burn-in under prolonged electrical and thermal stress. Existing data-driven compensation methods struggle to accurately compensate for uneven aging caused by localized heating under static or high-frequency drive conditions， as they neglect the spatial thermal coupling effects between pixels. To address these issues， this paper proposes a brightness compensation architecture based on thermal diffusion mechanisms. By integrating a discrete two-dimensional thermal diffusion model to approximate heat conduction between neighboring pixels， we designed a physically enhanced data simulation method capable of generating spatially correlated aging patterns. Based on this modeling architecture， we constructed a spatially aware deep neural network that employs a two-dimensional convolutional architecture combined with an attention mechanism to simultaneously capture local structural features and parameters sensitive to degradation. Experimental results show that， compared to traditional pixel-independent models， the proposed method improves brightness degradation accuracy and structural similarity by 2.7% and 0.273， respectively， under specified conditions—particularly in scenarios involving long-term aging and spatially non-uniform degradation. Qualitative visual comparisons confirm that the proposed method effectively suppresses burn-in artifacts while maintaining structural consistency in complex regions.关键词：brightness uniformity;self-attention mechanism;OLED;data augmentation8|0|0更新时间：2026-10-16
Display Technology and Applications
- 摘要：Driving parameters established for conventional electrowetting displays cannot readily be applied directly to devices incorporating quantum-dot-containing oil， and aperture retention during sustained driving requires improvement. This study compares different voltage-loading protocols， investigates the effects of bipolar square-wave frequency and positive-polarity duty cycle on aperture response， and analyzes luminance responses in comparison with those of conventional electrowetting devices. A startup-maintenance two-stage driving waveform is then developed. The results show that both frequency and positive-polarity duty cycle affect aperture formation in the quantum-dot devices， whose luminance-duty-cycle relationship differs markedly from that of conventional devices. Among the tested schemes， applying ±22 V for the first 0.5 s followed by ±18 V during the maintenance stage yields an average aperture ratio of 73.92% at 30 s and reduces the residual area ratio from 7.84% under the reference square wave to 3.48%. At a positive-polarity duty cycle of 50%， luminance increases by 8.20% under ambient light and by 20.95% under ambient light with additional 395 nm ultraviolet illumination， relative to the reference waveform. Matching the startup and maintenance voltage amplitudes to the driving-response characteristics of devices incorporating quantum dots can maintain an effective aperture while reducing residual oil and improving optical output. These findings provide an experimental basis for the targeted design and optimization of driving waveforms for quantum-dot electrowetting display devices.关键词：quantum-dot electrowetting display;bipolar driving;positive-polarity duty cycle;oil film splitting;two-stage driving3|0|0更新时间：2026-10-16
Driving and Controlling
- 摘要：To address the issues of insufficient frequency information utilization and the limited capability to represent high-frequency details in complex distortion scenarios for no-reference image quality assessment（NR-IQA）， a frequency-aware quality assessment method guided by human visual system （HVS） characteristics is proposed. First， a lightweight hierarchical Vision Transformer network is employed as the backbone for multi-scale feature extraction， enabling information modeling from local textures to global semantics. Second， a frequency-aware gating module inspired by the contrast sensitivity function（CSF） is introduced. This module adaptively modulates distortion features across channel， frequency， and spatial dimensions to enhance the model’s perceptual sensitivity to sensitive frequency information. Subsequently， a high-frequency enhancement module based on the Discrete Cosine Transform （DCT） is designed， which strengthens the model’s response to local high-frequency degradation through spatial and channel interactions in the frequency domain. Finally， multi-level features are aggregated to improve the collaborative representation capability among features. Experimental results demonstrate that the proposed method achieves a Pearson Linear Correlation Coefficient（PLCC） of 0.981 and a Spearman Rank-order Correlation Coefficient（SRCC） of 0.980 on the LIVE dataset. On the authentic distortion dataset LIVEC， the PLCC and SRCC reach 0.911 and 0.888， respectively， while the SRCC in the cross-dataset evaluation reaches 0.907. The proposed method effectively improves quality prediction accuracy in complex distortion scenarios， exhibits strong generalization and robustness， and can provide reliable support for perceptual quality optimization in intelligent visual systems.关键词：No-reference image quality assessment;contrast sensitivity function;frequency perception;discrete cosine transform;feature fusion2|0|0更新时间：2026-10-16
- 摘要：To address the issues of low contrast and detail loss in low-light images， as well as the output saturation and limited noise energy utilization of classical bistable stochastic resonance（SR） systems， this paper proposes an unsaturated SR method for low-light image enhancement based on Gaussian radial basis functions. An unsaturated symmetric bistable potential function composed of Gaussian radial basis functions and a linear term is constructed. An enhancement framework is designed， comprising subthreshold input signal generation， pixel-wise SR response computation， and response-guided brightness fusion. A brightness-darkness resonance separation metric is proposed to quantify the differential enhancement capability of SR， and a Comprehensive Enhanced Quality-Index（CEQI） is constructed from eight dimensions. Experimental results demonstrate that the proposed unsaturated potential function improves the resonance separation and expands the effective noise parameter range compared with the classical bistable potential. On low-light images with subtle and significant brightness contrasts， the proposed method achieves average CEQI values of 0.801 2 and 0.757 1， respectively， outperforming classical SR （0.704 3 and 0.692 9） as well as CLAHE， Gamma correction， MSRCR， and SSR. In real-life scene images， the average CEQI reaches 0.78. This study introduces unsaturated SR into low-light image enhancement， providing a new approach for recovering dark-region details while suppressing overexposure.关键词：stochastic resonance;image enhancement;Low-Light image;unsaturated potential function;gaussian radial basis function6|0|0更新时间：2026-10-16
- 摘要：To address the issues of small apparent size， weak edge features， and strong background interference of some road-surface foreign objects in complex road scenes captured by forward-facing vehicle-mounted cameras， a lightweight YOLO11n-based detection algorithm， YOLO11-MSDF， is proposed. The WFD-Down module decouples high- and low-frequency features to preserve edge and texture details. The DSA-SPPF module employs strip pooling to suppress structured background interference. The MAK-C3 module introduces adaptive sampling to enhance irregular-boundary features. The HADF module fuses shallow-layer details with deep-layer semantic information. Experimental results show that the model contains 2.9 million parameters and requires 7.5 GFLOPs， achieving a precision of 85.8%， a recall of 73.9%， and an mAP@0.5 of 83.7%， which are 7.1， 2.0， and 4.0 percentage points higher than those of YOLO11n， respectively. The detection speed reaches 90.5 FPS on the experimental platform. On an external dataset， the three metrics improve by 5.2， 1.3， and 1.9 percentage points， respectively. Experiments under synthetic fog and motion blur show that the proposed model outperforms the baseline under most degraded conditions， while the analysis of vehicle operating constraints identifies the main factors affecting practical vehicle-mounted applications.关键词：deep learning;road-surface foreign object detection;lightweight model;feature fusion;object detection7|0|0更新时间：2026-10-16
- 摘要：Printed Circuit Boards（PCBs） are fundamental components in electronic devices， and surface defects can affect the reliability and lifespan of electronic products. To address the issues of high computational cost， high storage overhead， and limited power consumption in edge deployment of PCB defect detection models， this paper designs a PCB defect detection system based on the ZYNQ SoC platform and the YOLO26n target detection model. To overcome the bottleneck of convolutional computation， this paper employs a Vitis HLS convolution acceleration IP core and combines spatial partitioning， channel parallelism， loop unrolling， and double-buffered ping-pong scheduling strategies to improve computational parallelism and data transmission efficiency. Simultaneously， 16-bit Q7.9 fixed-point quantization and extended bit-width accumulation are used to reduce data storage and transmission overhead. Experimental results show that all 170 operators in the network achieve floating-point and fixed-point path alignment， and the minimum cosine similarity between the convolutional topology and the Q7.9 fixed-point path are 0.989 4 and 0.999 8， respectively. CPU floating-point and ZYNQ fixed-point inference maintained similar performance across various defect detection metrics， with the largest differences in Precision， Recall， and mAP@0.5 being 0.023， 0.024， and 0.012， respectively. Hardware resource utilization rates of Look-Up-Table （LUT）， Flip-Flop （FF）， and Digital signal Processor （DSP） are 15.42%， 7.30%， and 20.64%， respectively， with an overall power consumption of 2.615 W. This indicates that the system can achieve low-power edge inference while maintaining high detection consistency.关键词：PCB defect detection;Zynq;YOLO26n;hardware acceleration7|2|0更新时间：2026-10-16
- 摘要：To address the issue of declining local evidence quality in Internet of Vehicles collaborative perception， which affects the stability and usability of collaborative fusion results， the fusion process is modeled as local perception evidence reliability estimation， and a spatiotemporal consistency-based robust fusion method（STC-RF） is proposed for collaborative perception. First， low-dimensional semantic evidence is constructed using a target existence heatmap. Subsequently， by comprehensively considering cross-vehicle spatial consistency， historical occupancy continuity， and negative evidence conflict， Bird’s-Eye View（BEV） grid-level local evidence reliability is generated and injected as a bias into the intermediate fusion of attention to continuously adjust the fusion contributions of different collaborators at various spatial positions. Furthermore， a conservative historical occupancy update is adopted to ensure that the historical time reference continuously participates in the fusion process， reducing the impact of local evidence fluctuations on the temporal continuity of the collaborative results. Experimental results demonstrate that under the condition of background evidence degradation， the proposed method improves AP@0.7 from 82.09% to 86.20%， reduces SSR from 8.21% to 5.53%， and decreases CFNR from 0.90% to 0.72%， indicating that the method enhances the fusion stability under non-ideal collaborative perception conditions.关键词：internet of vehicles;collaborative perception;perception evidence degradation;spatiotemporal consistency;robust fusion2|0|0更新时间：2026-10-16
- 摘要：Computed Tomography（CT） images of pulmonary tuberculosis suffer from issues including noise interference， weak lesion features， and insufficient utilization of three-dimensional spatial information， which impair classification performance. To address these problems， this paper proposes a three-class classification method for pulmonary tuberculosis CT images that integrates image enhancement and multi-dimensional feature learning. This method adopts a joint preprocessing strategy combining multi-scale eight-direction anisotropic diffusion enhancement and progressive N2F （Noise2Fast） denoising， which improves the visibility of lesion textures. Subsequently， a lightweight Parallel Representation with PRISM-Net（Iterative Spatial Multi-view Fusion Network） is constructed. The network utilizes an Iterative Multi-view Fusion（IMVF） strategy to extract complementary information from different spatial orientations， and introduces a MDAM （Multi-Dimensional Attention Module） to enhance the expression capability of three-dimensional features. Validated on an independent test set comprising 150 cases， the proposed method achieves a precision of 94.68% and a weighted F1-score of 94.63%. The AUC （Area Under the Curve） for normal samples， active pulmonary tuberculosis， and inactive pulmonary tuberculosis reaches 1.000 0， 0.992 6， and 0.992 3 respectively， outperforming benchmark models including MambaOut-Kobe， MobileNetV3， MedVit and ResNet50.关键词：pulmonary tuberculosis classification;image enhancement;multi-view fusion;multi-dimensional attention mechanism;three-dimensional convolutional neural network7|3|0更新时间：2026-10-16
-
Image Processing
- Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China 130033 Postal code：130033
- Tel：0431-86176059 Email：yjxs@ciomp.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0