A novel computer-aided detection method based on deep learning framework was proposed to detect small intestinal ulcer and erosion in wireless capsule endoscopy (WCE) images. To the best of our knowledge, this is the first time that deep learning framework has been exploited on automated ulcer and erosion detection in WCE images. Compared with the traditional detection method, deep learning framework can produce image features directly from the data and increase recognition accuracy as well as efficiency, especially for big data. The developed method included image cropping and image compression. The AlexNet convolutional neural network was trained to the database with tens of thousands of WCE images to differentiate lesion and normal tissue. The results of ulcer and erosion detection reached a high accuracy of 95.16% and 95.34%, sensitivity of 96.80% and 93.67%, and specificity of 94.79% and 95.98%, correspondingly. The area under the receiver operating characteristic curve was over 0.98 in both of the networks. The promising results indicate that the proposed method has the potential to work in tandem with doctors to efficiently detect intestinal ulcer and erosion.
When surface plasmon resonance (SPR) occurs, the incident light is absorbed by the surface of the SPR structure, thus minimizing the intensity of the reflected light. Therefore, the SPR method is adopted in this paper to achieve perfect absorption of the absorbent. In this paper, we first propose a multi-frequency broadband absorber structure based on graphene SPR, which uses the continuous resonance of patterned graphene surface plasmon in the frequency spectrum to form a multi-frequency broadband absorption. In this simulation, a sandwich-stack structure was adopted, whereby the patterned graphene is situated on top of the SiO2 layer and the metal layer. The broad-band absorption bands of the absorber were obtained as 4.14–4.38 THz, 5.78–6.36 THz, and 7.87–8.66 THz through the analog simulation of finite-difference time-domain method (FDTD) solutions. Then, based on the multi-layer resonant unit structure, through the superposition and combination of absorbing units responding to different frequency bands, the perfect absorption of ultra-wideband is achieved. The data results illustrate that the total absorption bandwidth of the absorber is 2.26 THz, and the relative absorption bandwidth Bw is equal to 28.93%. The electric field in X-Y direction of the absorber in the perfect absorption band is analyzed, respectively, and the dynamic tunability of the absorber is studied. Finally, we studied whether the absorbing structure still has efficient absorption characteristics for the two polarization modes when the incident angle is changed from 0° to 70°. The structure model proposed has potential value for application in terahertz photoelectric detection, filtering, and electromagnetic shielding.
A three-dimensional (3D) photonic nanojet (PNJ) emerging from a liquid-immersed core–shell dielectric microsphere is numerically investigated by the finite-difference time-domain (FDTD) method. An ultra-elongated PNJ with an effective length larger than 57 wavelengths while retaining a high intensity and a large working distance is obtained from the simulation. In particular, PNJ properties, including intensity enhancement, working distance, effective length, and full width at half maximum (FWHM), can be well tuned and controlled by varying the refractive index of the immersed liquid. We believe that this design is applicable to many fields, such as material science, nanophotonics, and biomedicine.
We report a novel quasi-synchronously pumped PPMgLN-based high power mid-infrared (MIR) laser with picosecond pulse bunch output. The pump laser is a linearly polarized MOPA structured all fiberized Yb fiber laser with picosecond pulse bunch output. The output from a mode-locked seed fiber laser was directed to pass through a FBG reflector via a circulator to narrow the pulse duration from 800 ps to less than 50 ps and the spectral FWHM from 9 nm to 0.15 nm. The narrowed pulses were further directed to pass through a novel pulse multiplier through which each pulse was made to become a pulse bunch composing of 13 sub-pulses with pulse to pulse time interval of 1.26 ns. The pulses were then amplified via two stage Yb fiber amplifiers to obtain a linearly polarized high average power output up to 85 W, which were then directed to pass through an isolator and to pump a PPMgLN-based optical parametric oscillator via quasi-synchronization pump scheme for ps pulse bunch MIR output. High MIR output with average power up to 4 W was obtained at 3.45 micron showing the feasibility of such pump scheme for ps pulse bunch MIR output.
A multi-channel light emitting diode (LED)-induced fluorescence system combined with a convolutional neural network (CNN) analytical method was proposed to classify the varieties of tea leaves. The fluorescence system was developed employing seven LEDs with spectra ranging from ultra-violet (UV) to blue as excitation light sources. The LEDs were lit up sequentially to induce a respective fluorescence spectrum, and their ability to excite fluorescence from components in tea leaves were investigated. All the spectral data were merged together to form a two-dimensional matrix and processed by a CNN model, which is famous for its strong ability in pattern recognition. Principal component analysis combined with k-nearest-neighbor classification was also employed as a baseline for comparison. Six grades of green tea, two types of black tea and one kind of white tea were verified. The result proved a significant improvement in accuracy and showed that the proposed system and methodology provides a fast, compact and robust approach for tea classification.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.