
Search by keywords or author
Export citation format
Architectural style classification algorithm fusing CNN and Transformer
Dong LIU, Rongfu ZHANG, Junxiang QIN, Junzhe GONG, and Zhibin CAO
The accurate classification of architectural style is of great significance to the study of architectural culture and human history and civilization. Models based on convolutional neural network (CNN) has achieved good performance in the field of architectural style classification due to its powerful feature extractionThe accurate classification of architectural style is of great significance to the study of architectural culture and human history and civilization. Models based on convolutional neural network (CNN) has achieved good performance in the field of architectural style classification due to its powerful feature extraction ability. However, most current CNN models only extract the local features of architecture buildings. With the attention mechanism, a model based on Transformer can extract the globle features of architecture buildings. In order to improve the accuracy of architectural style classification, an architectural style classification method fusing CNN and Transformer is proposed. The core of the network is CT-Block structure. In terms of channel dimension, the structure is divided into two branches, CNN and Transformer, and the features pass through the two channels respectively and then concatenate together. This structure then concatenate together. This structure can not only fuse the local features extracted by CNN and the global features extracted by Transformer, but also alleviate the problem of model size and parameter number increase caused by the two-branch structure. The experimental results of Architectural Style Dataset and WikiChurches dataset were 79.83% and 68.41% respectively, which was better than other models in the field of architectural style classification..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 1 (2024)
Retinal blood vessel segmentation algorithm based on improved U-Net
Yuan LIU, Baicheng LI, and Chunbo WU
The existing algorithms have the problem of low segmentation accuracy when facing small vessels in retinal images. In this paper, an improved U-Net segmentation algorithm is proposed by introducing residual module and detail enhancement attention mechanism module into U-Net network. In the coding and decoding stages, tThe existing algorithms have the problem of low segmentation accuracy when facing small vessels in retinal images. In this paper, an improved U-Net segmentation algorithm is proposed by introducing residual module and detail enhancement attention mechanism module into U-Net network. In the coding and decoding stages, the residual module is used to replace the traditional convolutional module, which solves the problem of network degradation with increasing depth. Meanwhile, a detail enhancement attention mechanism is added between the encoder and the decoder to reduce the useless information in the output of the encoder, so that the sensitivity of the network to capture valid feature information is improved. In addition, the experimental results based on the standard image set DRIVE reveal that the segmentation accuracy, sensitivity and F1 score of the proposed algorithm are improved by 0.46%, 2.14% and 1.56%, respectively compared to the U-Net, which is superior to the traditional segmentation algorithms..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 9 (2024)
Home-based pancreatic adenocarcinoma test chip based on microfluidic chip and immunofluorescence chromatography
Jingwei CHEN, and Lulu ZHENG
The high incidence and mortality of pancreatic adenocarcinoma have attracted widespread attention in recent years. However, the existing detection methods are costly and complex to operate, which makes it difficult to extend their use to households in the community. On the other hand, the lower-cost traditional immunocThe high incidence and mortality of pancreatic adenocarcinoma have attracted widespread attention in recent years. However, the existing detection methods are costly and complex to operate, which makes it difficult to extend their use to households in the community. On the other hand, the lower-cost traditional immunochromatographic test strips are difficult to detect the markers in blood samples effectively. A microfluidic-immunochromatographic test strip for tumor marker CA199 was designed. With the aid of the optical microscope, the microarray structure could effectively separate red blood cells in sample solution without destroying the cellular structure. The physical properties of the quantum dot fluorescent microspheres were analyzed by a transmission electron microscope and a multifunctional enzyme marker. The study shows that the quantum dot fluorescent microspheres have good stability and dispersity. By comparison with unfiltered samples, the enhancement of microfluidic structures for CA199 fluorescent signals in samples was demonstrated. The detection limit of the system is 4.2 U/mL. The microfluidic-immunochromatographic test strips are low-cost, simple-to-operate, safe and stable. It provides a new strategy for home-based screening for pancreatic adenocarcinoma..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 17 (2024)
Image processing method for polarization detection of microdefects in optical components
Xiaohui CHEN, Xinsen WANG, and Yafan DUAN
Surface defects of optical components seriously affect the performance of optical systems. The existing surface defects detection technology of optical components has problems such as slow detection speed and low accuracy. This paper studies the digital image processing method of polarization imaging of defects based oSurface defects of optical components seriously affect the performance of optical systems. The existing surface defects detection technology of optical components has problems such as slow detection speed and low accuracy. This paper studies the digital image processing method of polarization imaging of defects based on the polarization detection technology of micro-defects of optical components. Firstly, we used the polarization property of light to capture the image with high contrast of defects, and then pre-processed the image by filtering and denoising, threshold segmentation, morphological processing. Next, different feature parameters are selected according to different types of defects so as to achieve defect detection and classification, meanwhile accurately measuring the size of defects. The results showed that the total defect detection rate of this method was 95.90%, and the detection time of a single image was less than 50 ms..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 24 (2024)
Design of ATE data storage and transmission system for optical chip testing
Qifan JIA, Xuanhong JIN, Pengcheng XIAO, and Hangyu HE
Optical chip is an important device to realize photoelectric signal conversion in optical fiber system, so it needs to undergo a lot of tests after manufacturing.As the main equipment used in chip testing today, integrated circuit automatic test equipment (ATE) has the advantages of high automation, high precision, widOptical chip is an important device to realize photoelectric signal conversion in optical fiber system, so it needs to undergo a lot of tests after manufacturing.As the main equipment used in chip testing today, integrated circuit automatic test equipment (ATE) has the advantages of high automation, high precision, wide test range and fast speed in the test of optical chip. Since the ATE device generates a great many data each time it is tested, and a long test time will increase the test cost, so the ATE system requires a high data rate. In order to improve the Data transmission efficiency and storage during optical chip testing, a high-speed data storage and transmission system based on embedded system chip and FPGA was designed. The enhanced direct memory access module was used to drive the GPMC interface to improve the transmission bandwidth, and the data transmission mode was optimized on FPGA to improve the DDR3 SDRAM utilization. After testing on the board, the read and write bandwidth in single channel reaches 413.3 Mbit/s and 984.6 Mbit/s respectively, realizing high-speed transmission and stable storage of data in the system..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 31 (2024)
The 465 nm light emission based on all-inorganic perovskite quantum dots
Liyun XU, and Gongjie XU
The Rec.2020 blue light-emitting standard requires the center wavelength is (465 ± 3) nm, and the full width at half maxima (FWHM) is less than 20 nm, and the color gamut coordinate is (0.131, 0.046). To benchmark the standard, the performance of all inorganic perovskite quantum dots was investigated. The various The Rec.2020 blue light-emitting standard requires the center wavelength is (465 ± 3) nm, and the full width at half maxima (FWHM) is less than 20 nm, and the color gamut coordinate is (0.131, 0.046). To benchmark the standard, the performance of all inorganic perovskite quantum dots was investigated. The various blue light-emitting series were synthesized via the methods of adjusting halogen ratio, ion-exchange, and adjusting the number of octahedral layers, respectively. Their performances for Rec. 2020 blue light-emitting standard were discussed comprehensively. The results indicate that the CsPbCl1.5Br1.5 obtained by adjusting the halogen ratio has the best performance in terms of light-emitting wavelength, monochromaticity, luminescence efficiency, etc., with a center wavelength of 463.9 nm, an FWHM of 18.6 nm, and a color gamut coordinate of (0.137, 0.041). This work lays a foundation for the further development of high-definition blue light emitting devices and display systems based on all-inorganic perovskite systems..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 40 (2024)
Multi-modal image reconstruction method based on Trans-MIR model
Yiming LI, Hao WANG, Ran LI, Quan CHEN, Haijun LU, and Hui YANG
Image reconstruction is one of the key steps in the optical computational imaging. At present, image reconstruction based on deep learning mainly uses convolutional neural network, cyclic neural network and generative adversarial network. Most models are only trained through the data of a single mode, which is difficulImage reconstruction is one of the key steps in the optical computational imaging. At present, image reconstruction based on deep learning mainly uses convolutional neural network, cyclic neural network and generative adversarial network. Most models are only trained through the data of a single mode, which is difficult to ensure the quality of imaging while possessing the generalization ability of different scenes. To solve this problem, a multi-modal image reconstruction model based on the Transformer (Trans-MIR) is proposed in this paper. Experimental results show that Trans-MIR can extract image features from multi-modal data to achieve high-quality image reconstruction. The structural similarity of 2D universal face speckle reconstruction was as high as 0.93 and the mean square error of 3D microtubule reconstruction was as low as 10-4. It provides inspiration for the study of multimodal image reconstruction..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 51 (2024)
Theoretical simulation of terahertz near-field scanning microscopic imaging
Li ZHOU, Xitian HU, and Xu WU
Terahertz near-field scanning imaging system breaks through the limitation of the diffraction limit of the traditional far-field optical system and can achieve super-resolution detection and imaging in terahertz band. Based on the working principle of a scattering-type terahertz scanning near-field optical microscope, Terahertz near-field scanning imaging system breaks through the limitation of the diffraction limit of the traditional far-field optical system and can achieve super-resolution detection and imaging in terahertz band. Based on the working principle of a scattering-type terahertz scanning near-field optical microscope, we established a numerical simulation model of a scattering-type terahertz scanning near-field optical microscope based on COMSOL and verified lightning rod effect and antenna resonance effect generated by the probe excited by incident light, as well as the dipole effect between the probe and the sample. On this basis, we implemented near-field signal calculation and near-field scanning imaging of protein molecules with reference to actual scanning processes. The results show that the model can achieve near-field imaging of three-dimensional materials and has application potential in the study of complex three-dimensional micro nano materials..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 58 (2024)
Low light image enhancement based on semantic information and attention mechanism
Haobin LI, and Yunsong HUA
Aiming at the problems of low contrast and noise in low-light images, this paper proposes a low light enhancement method which combines semantic information and attention mechanism. First, a pair of jointly trained U-Net networks were used to obtain the preliminary enhancement results and the distribution probability oAiming at the problems of low contrast and noise in low-light images, this paper proposes a low light enhancement method which combines semantic information and attention mechanism. First, a pair of jointly trained U-Net networks were used to obtain the preliminary enhancement results and the distribution probability of semantic information of low-light images by sharing feature extractors. Then, the low-light enhancement features and semantic features obtained by U-Net networks were fused through the attention mechanism module. The problem of image edge information loss under low illumination and image blurring under exposure was addressed. Experiment results show that the proposed method can effectively eliminate artifacts when processing low illumination images with low contrast and uneven exposure, and improve image saturation and contrast of different regions..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 65 (2024)
Pressure measurement method and device using magnetic balance
Binbin TU, Fei CAI, Jufeng ZHAO, and Guangmang CUI
In piston pressure measurement, small pressure changes rely on small weights, which cannot achieve continuous change of pressure output, and weights-based pressure measurement also has difficulties in automation. This paper proposes a pressure measurement device using magnetic balance based on opto-mechanical design, tIn piston pressure measurement, small pressure changes rely on small weights, which cannot achieve continuous change of pressure output, and weights-based pressure measurement also has difficulties in automation. This paper proposes a pressure measurement device using magnetic balance based on opto-mechanical design, to explore the feasibility of the electromagnetic balance method to achieve pressure measurement. In the device, the weight gravity measurement originally used for balance is replaced by an electronic scale, without manual weight addition and subtraction, and the balance force is directly weighed by the scale. Through magnetic repulsion, the spacing in the magnet pair is adjusted to continuously output the required pressure. This method can transform the process of manually placing weights for the balance into the process of real-time continuous force value reading of the scale to finally achieve balance. The experimental analysis demonstrates the influence of the magnet on the reading of the electronic scale. In the pressure output test experiment and analysis, the precise relationship of the distance between the magnets and the magnetic force output is obtained, which proves that the device is effective. It is feasible to replace the small weight in piston manometer with the electromagnetic balance method..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 75 (2024)
Polarization detection device based on catenary metasurface
Shanhu GAO, Mingzhu XU, Long CHEN, and Xiumin GAO
Traditional optical devices cannot meet the requirements of modern optical integration because of their bulky size and sensitivity to linearly polarized light or circularly polarized light. In this paper, based on surface plasmon polaritons (SPPs), a universal polarization analyzer consisting of catenary was designed. Traditional optical devices cannot meet the requirements of modern optical integration because of their bulky size and sensitivity to linearly polarized light or circularly polarized light. In this paper, based on surface plasmon polaritons (SPPs), a universal polarization analyzer consisting of catenary was designed. The theory of SPPs propagation was derived, and the electric field pattern of catenary and the distribution of electric field intensity in the center of the device were calculated by using the finite difference time domain method. The results show that when light of different polarization states are incident separately, catenary has a variety of electric field patterns. The significantly differentiated electric field distribution in the center of the device can effectively distinguish the polarization states of linearly polarized light, circular polarized light and vector polarized light (radial polarized light and angular polarized light). Therefore, these results are helpful for the research and development of the catenary multifunctional devices..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 81 (2024)
Study on anti-radiation damage effect of MgF2/Lumogen composite films
Ruicheng JIANG, Yanlei LYU, Shijie LIU, and Chunxian TAO
In order to investigate the characteristics of high-energy radiation damage in MgF2/Lumogen composite films used for vacuum ultraviolet detection, a 300 W high-energy continuous laser with a wavelength of 1064 nm was employed for irradiation. By adjusting the laser power and scanning rate, the energy and duration of laIn order to investigate the characteristics of high-energy radiation damage in MgF2/Lumogen composite films used for vacuum ultraviolet detection, a 300 W high-energy continuous laser with a wavelength of 1064 nm was employed for irradiation. By adjusting the laser power and scanning rate, the energy and duration of laser irradiation on the film were controlled, and measurements were conducted on photoluminescence intensity, fluorescence decay, and transmittance before and after irradiation. The results demonstrate that under laser irradiation with an energy density ranging from 0.58 to 2.94 W/mm2 at a scanning rate between 50 to 10 mm/min, the fluorescence response intensity upon deep ultraviolet excitation at 160 nm decreases by approximately 1% to 61%, while that upon near-ultraviolet excitation at 273 nm decreases by around 1% to 21%. Moreover, there is a decrease in visible light transmittance within the range of wavelengths from 500 to700 nm by about 2% to 3%. Based on the comparison of the radiation damage observed in single-layer Lumogen films as well as heat conduction analysis performed on composite films, it can be concluded that MgF2 film not only possesses anti-radiation damage capability but also exhibits an anti-reflection effect on Lumogen..
Optical Instruments
- Publication Date: Oct. 25, 2024
- Vol. 46, Issue 5, 88 (2024)