Catalogue Search | MBRL
Are you sure you want to remove the book from the shelf?
{{itemTitle}}
446
result(s) for
"Cheng-Fu, Yang"
Sort by:
غبار تحت الشمس : قصص قصيرة
يعرض الكتاب بين دفتيه قصصا قصيرة لعادات الأقليات الصينية، وسلوكيات العائلات في المناسبات الدينية والاحتفالات الأخرى المختلفة، وتتباين فيها الشخصيات والزمان والمكان، لتضع القارئ أمام رسم أدبي يشرح حياة الأقليات الصينية المليئة بالتفاصيل والمفارقات. الكتاب من تأليف: يه شنغ فو، وترجمة عن الصينية.
Book
Towards a holistic framework for multimodal LLM in 3D brain CT radiology report generation
Multi-modal large language models (MLLMs) have transformed the landscape of modern healthcare, with automated radiology report generation (RRG) emerging as a cutting-edge application. While 2D MLLM-based RRG has been well established, its utility for 3D medical images remains largely unexplored. In this regard, we curate the 3D-BrainCT dataset (18,885 text-scan pairs) and develop BrainGPT, a clinically visual instruction-tuned (CVIT) model designed for 3D CT RRG. While we notice that the traditional LLM metrics failed to gauge the diagnostic quality of the RRG, we propose feature-oriented radiology task evaluation (FORTE), an evaluation scheme that captures the clinical essence of the generated reports. Here we show that BrainGPT achieves an average FORTE F1-score of 0.71 (degree = 0.661; landmark = 0.706; feature = 0.693, and impression = 0.779) and 74% of BrainGPT-generated reports were indistinguishable from human-written ground truth in a
Turing
-like test. Together, our work establishes a comprehensive framework encompassing dataset curation, anatomy-aware model fine-tuning, and the development of robust evaluation metrics for the RRG. By sharing our experience in 3D MLLM-based RRG, we aim to accelerate the expedition in human-machine collaboration for next-generation healthcare.
Multimodal large language models (MLLMs) hold promise for a range of medical applications. Here, the authors use MLLMs for 3D brain CT radiology report generation, demonstrating that combining anatomy-aware model fine-tuning with robust evaluation metrics establishes a comprehensive and effective framework.
Journal Article
Research and Evaluation on an Optical Automatic Detection System for the Defects of the Manufactured Paper Cups
In this paper, the paper cups were used as the research objects, and the machine vision detection technology was combined with different image processing techniques to investigate a non-contact optical automatic detection system to identify the defects of the manufactured paper cups. The combined ring light was used as the light source, an infrared (IR) LED matrix panel was used to provide the IR light to constantly highlight the outer edges of the detected objects, and a multi-grid pixel array was used as the image sensor. The image processing techniques, including the Gaussian filter, Sobel operator, Binarization process, and connected component, were used to enhance the inspection and recognition of the defects existing in the produced paper cups. There were three different detection processes for paper cups, which were divided into internal, external, and bottom image acquisition processes. The present study demonstrated that all the detection processes could clearly detect the surface defect features of the manufactured paper cups, such as dirt, burrs, holes, and uneven thickness. Our study also revealed that the average time for the investigated Automatic Optical Detection to detect the defects on the paper cups was only 0.3 s.
Journal Article
Optimized YOLOv3 Algorithm and Its Application in Traffic Flow Detections
In the intelligent traffic system, real-time and accurate detections of vehicles in images and video data are very important and challenging work. Especially in situations with complex scenes, different models, and high density, it is difficult to accurately locate and classify these vehicles during traffic flows. Therefore, we propose a single-stage deep neural network YOLOv3-DL, which is based on the Tensorflow framework to improve this problem. The network structure is optimized by introducing the idea of spatial pyramid pooling, then the loss function is redefined, and a weight regularization method is introduced, for that, the real-time detections and statistics of traffic flows can be implemented effectively. The optimization algorithm we use is the DL-CAR data set for end-to-end network training and experiments with data sets under different scenarios and weathers. The analyses of experimental data show that the optimized algorithm can improve the vehicles’ detection accuracy on the test set by 3.86%. Experiments on test sets in different environments have improved the detection accuracy rate by 4.53%, indicating that the algorithm has high robustness. At the same time, the detection accuracy and speed of the investigated algorithm are higher than other algorithms, indicating that the algorithm has higher detection performance.
Journal Article
Using Screen Printing Technology to Fabricate Flexible Sodium Ion Sensors
This study focused on the development of Na+ ion sensing devices on a flexible substrate and investigated the impact of various additive materials on its sensing performance. For the Na+ ion sensing aspect, the film on the carbon working electrode used tert-butyl calix[4]arene tetraethyl acetate as the ion carrier. The main component of the film was polyvinyl chloride (PVC), with a plasticizer added to enhance its flexibility, ensuring better adaptation to the flexible substrate. In this base formulation, graphene oxide (GO) or multi-walled carbon nanotubes (MWCNTs) were incorporated into the sensing electrode to explore their effects on Na+ ion sensing capabilities. The results demonstrated that adding MWCNTs significantly improved the sensor’s sensitivity to Na+ ions. In addition, the study used the response slope to Na+ ions as a comparative reference for selectivity by calculating the ratio of the Na+ ion response slope to the response slopes of other ions (such as K+ and Ca2+). The findings showed that the sensors with MWCNTs exhibited better selectivity than the others with GO, and therefore, further analysis was performed on the response time of the sensors with MWCNTs. The results indicated that incorporating MWCNTs reduced the sensors’ response time and enhanced their overall sensitivity. However, excessive addition of MWCNTs would lead to a decrease in the selectivity of the fabricated sensors. This suggests that while MWCNTs offer promising improvements in performance, their concentration must be carefully optimized to maintain the sensors’ selectivity.
Journal Article
Fabrications of the Flexible Non-Enzymatic Glucose Sensors Using Au-CuO-rGO and Au-CuO-rGO-MWCNTs Nanocomposites as Carriers
A flexible, non-enzymatic glucose sensor was developed and tested on a polyethylene terephthalate (PET) substrate. The sensor’s design involved printing Ag (silver) as the electrode and utilizing mixtures of either gold–copper oxide-modified reduced graphene oxide (Au-CuO-rGO) or gold–copper oxide-modified reduced graphene oxide-multi-walled carbon nanotubes (Au-CuO-rGO-MWCNTs) as the carrier materials. A one-pot synthesis method was employed to create a nanocomposite material, consisting of Au-CuO-rGO mixtures, which was then printed onto pre-prepared flexible electrodes. The impact of different weight ratios of MWCNTs (0~75 wt%) as a substitute for rGO was also investigated on the sensing characteristics of Au-CuO-rGO-MWCNTs glucose sensors. The fabricated electrodes underwent various material analyses, and their sensing properties for glucose in a glucose solution were measured using linear sweep voltammetry (LSV). The LSV measurement results showed that increasing the proportion of MWCNTs improved the sensor’s sensitivity for detecting low concentrations of glucose. However, it also led to a significant decrease in the upper detection limit for high-glucose concentrations. Remarkably, the research findings revealed that the electrode containing 60 wt% MWCNTs demonstrated excellent sensitivity and stability in detecting low concentrations of glucose. At the lowest concentration of 0.1 μM glucose, the nanocomposites with 75 wt% MWCNTs showed the highest oxidation peak current, approximately 5.9 μA. On the other hand, the electrode without addition of MWCNTs displayed the highest detection limit (approximately 1 mM) and an oxidation peak current of about 8.1 μA at 1 mM of glucose concentration.
Journal Article
Effect of V2O5 B-site substitution on the microstructure, Raman spectrum, and dielectric properties of SrBi2Ta2O9 ceramics
Strontium bismuth tantalate vanadate [SrBi
2
(Ta
2−x
V
x
)O
9
, SBTV] ceramics, which are bismuth-layered perovskite ferroelectrics, were synthesized through the solid-state reaction method. The effects of different sintering temperatures and V
2
O
5
contents on the structure of the microstructure, Raman spectrum, and dielectric properties of the SBTV ceramics were investigated. As sintered at high temperature (980–1040 °C) and different V
2
O
5
contents (x = 0.1 − x = 0.4), only disk-like grains of the SBTV ceramics were observed in the scanning electron micrographs. Preferential orientation of the crystals of the SBTV ceramics was confirmed through X-ray diffraction studies. The higher dielectric constant and Curie temperature of the SBTV ceramics compared with those of strontium bismuth tantalite (SrBi
2
Ta
2
O
9
, SBT) ceramics are ascribe to the partial replace of Ta
5+
ions by V
5+
ions in the B sites. The Curie–Weiss law and the modified Curie–Weiss law were used to discuss the normal-type or relaxor-type ferroelectric characteristic of the SBTV ceramics. The Ta
5+
ion replaced by V
5+
ion site in SBT ceramics to form SBTV ceramics exerted a pronounced effect on the BO
6
mode, as demonstrated by Raman spectrum results.
Journal Article
Synthesis of ZnO Nanoflower Arrays on a Protrusion Sapphire Substrate and Application of Al-Decorated ZnO Nanoflower Matrix in Gas Sensors
In this study, we utilized a sapphire substrate with a matrix protrusion structure as a template. We employed a ZnO gel as a precursor and deposited it onto the substrate using the spin coating method. After undergoing six cycles of deposition and baking, a ZnO seed layer with a thickness of 170 nm was formed. Subsequently, we used a hydrothermal method to grow ZnO nanorods (NRs) on the aforementioned ZnO seed layer for different durations. ZnO NRs exhibited a uniform outward growth rate in various directions, resulting in a hexagonal and floral morphology when observed from above. This morphology was particularly evident in ZnO NRs synthesized for 30 and 45 min. Due to the protrusion structure of ZnO seed layer, the resulting ZnO nanorods (NRs) displayed a floral and matrix morphology on the protrusion ZnO seed layer. To further enhance their properties, we utilized Al nanomaterial to decorate the ZnO nanoflower matrix (NFM) using a deposition method. Subsequently, we fabricated devices using both undecorated and Al-decorated ZnO NFMs and deposited an upper electrode using an interdigital mask. We then compared the gas-sensing performance of these two types of sensors towards CO and H2 gases. The research findings indicate that sensors based on Al-decorated ZnO NFM exhibit superior gas-sensing properties compared to undecorated ZnO NFM for both CO and H2 gases. These Al-decorated sensors demonstrate faster response times and higher response rates during the sensing processes.
Journal Article
Publisher Correction: Effect of V2O5 B-site substitution on the microstructure, Raman spectrum, and dielectric properties of SrBi2Ta2O9 ceramics
An amendment to this paper has been published and can be accessed via a link at the top of the paper.
Journal Article
Morphological, Optical, and Electrical Properties of p-Type Nickel Oxide Thin Films by Nonvacuum Deposition
In this study, a p-type 2 at% lithium-doped nickel oxide (abbreviation L2NiO) solution was prepared using Ni(NO3)2·6H2O, and LiNO3·L2NiO thin films were deposited using an atomizer by spraying the L2NiO solution onto a glass substrate. The sprayed specimen was heated at a low temperature (140 °C) and annealed at different high temperatures and times. This method can reduce the evaporation ratio of the L2NiO solution, affording high-order nucleating points on the substrate. The L2NiO thin films were characterized by X-ray diffraction, scanning electron microscopy, UV–visible spectroscopy, and electrical properties. The figure of merit (FOM) for L2NiO thin films was calculated by Haacke’s formula, and the maximum value was found to be 5.3 × 10−6 Ω−1. FOM results revealed that the L2NiO thin films annealed at 600 °C for 3 h exhibited satisfactory optical and electrical characteristics for photoelectric device applications. Finally, a transparent heterojunction diode was successfully prepared using the L2NiO/indium tin oxide (ITO) structure. The current–voltage characteristics revealed that the transparent heterojunction diode exhibited rectifying properties, with a turn-on voltage of 1.04 V, a leakage current of 1.09 × 10−4 A/cm2 (at 1.1 V), and an ideality factor of n = 0.46.
Journal Article