機器學習

NCREE has originally developed 5D digital space—on the basis of 3D city modelsconnections of different kinds of sensors around the world— is an online to offline virtual space with a combination of rising 5G technology advantages. Collectingreorganizing various 3D cartographic data with Building Information Modeling (BIM), satellite imagery, UAV 3D modeling, LiDar point cloud data, etc., can increase the diversity of the buildinglandscape. 5D+5G smart city platform would accelerate the 5D smart city to become the digital twin of a real city.

1. CNN model architecture optimization techniques: Propose an Agilev3-L architecture, with a performance index of mAP@50:95FPS:FP16 up to 171.79 2. GOP-mode acceleration scheme for real time inference: employ GOP-modeTRACKING algorithms to enhance the processing rate. 3. Propose a rapid data labeling system HiTag.

The major cost of conventional RFID tags comes from a microchip. To reduce the start-up cost, we develop a chipless RFID system that eliminates the use of the chip. This research organize the real-world issues of chipless RFID into four subjects, including the design of chipless tags, chipless RFID readers, reader signal processing,wireless propagation channel. We successfully integrate the four topics, developing a calibration-free, orientation-insensitive,high-capacity chipless RFID system with multi-tag detection capability.

Using PVDF-HFP, lithium salt, LLZOplasticizer to prepare HSE membrane (ơi=4.6×10-4 S/cm, 25oC). The Free-standingcathode-supported membranes in the pouch-type NCM811HSELi ASSLBs were designed to enhance the mechanical strength (Li-Nafion-coated cellulose support, the total membrane thickness is 40-60 µm), improve Li-ion conductivity (Doped LLZO/Li salt), reduce interfacial impedance (PVDF-HFP/Li salt/plasticizer),inhibit Li dendritic deposition. The energy density of the ASSLBs has reached 300 Wh/kg, which exceeds the current best LIBs by 50.

The self-powered sensing system realizes self-powered wireless wearable pressure sensing modules, self-powered intelligent environment sensing robots,self-powered smart window, which can carry out self-powered sensing signals combined with the transmission of the Internet of Things. Functions such as medical monitoring aidsautomated environmental quality sensing. In the future, this system opens a window towards automatic monitoring of environmental analysispersonal smart healthcare.

Devices with physical flexibilitystretchability have attracted a great deal of interest for use in wearable electronic technologylarge-area electronics, including displays, energy harvesters, energy storage devices, distributed sensor networks,Internet of Things applications. The present invention relates to methods for growingtransferring single-crystal III-nitride nanorodstwo dimensional transition metal dichalcogenides (2D TMDs) nanostructures,devices having the transferred nanostructures for flexible device, microLED,3DIC fabrications.

R3CE (Rapid, Reproducible, Rare Cell 3D Expansion) is a clinically validated platform capable of generating single cell-derived organoids within one to four weeks. It outperforms currently 3D culture technologies in terms of yields, time to form organoids,easiness of operation. It also demonstrates similar yieldscost structure to 2D culture protocols. R3CE is ready for drug testing for all clinical samplesprovides sufficient quantity of genomic samples derived from cancer patients. It would help conduct clinical cancer treatment in the translational study.

"We present here a portable device with isothermally mediated loop amplification (LAMP) for the diagnosis of infectious diseases. The main components included (1) roller conveyor, (2) hotplate,(3) temperaturemotion control panels. After mixing all the solutions by rollingsqueezing the sealed compartments, the LAMP strip was heated for 30 min for amplification. Finally, fluorescence results can be read with the naked eyeon a smartphone under the UV lamp. The 700g device is powered by a portable charger, making it easier to detect pathogens in resource-limited environments."

The technology optimizes the antigen protein expression system, so that the structural protein Erns of CFSV converts insoluble proteins into soluble proteins during the production process. The breakthrough of this technology allows the antigen protein to be applied to the development of an immune attachment screening platform, as a specialized differential diagnostic reagents can distinguish between virus-infectedvaccinated pigs, which can be applied to the monitoring system of virus transmission rangevaccination

3D digital pathology imagingAI auxiliary diagnosis platform integrates 3D pathology sample preparation, high speed image scanning,intelligent software analysis into a novel system. We have collaborated with clinical pathologists to develop both 3D biopsy pathology imaging workflowthe world first 3D pathology image analysis software MetaLite, along with AI models pathology feature annotationquantitative analysis of solid tumor with high accuracy can be rapidly processed. We have transferred patented technology to start up JelloX Biotech Inc. at NTHU for further operation.

We proposed a positioning integrated gyroscope (PIG) with a unit price of less than US$3,000. A pilot test was conducted in the tunnel section. The positioning error of the MEMS gyroscope was 4,300 meters at a moving distance of 7.8 kilometers,because there was no GNSS signal in the tunnel. The signal of GNSS module only uses the built-in IMU for positioning, the maximum error is also 348 meters. We move 7.8 kilometers, the error is 89 meters73 meters, the overall error are MEMS (55), GNSS (4.5),the team＇s module (1.1), the internationally marketed inertial FOG (0.9).

Lung-on-minilab is designed as a tool for personalized healthcarewill contribute in precision medicine. Here, the tumor microenvironment was reconstituted using liquid dielectrophoresisarranged as lung tissue. The size of pattern can be customized as stages of cancer. Additionally, this chip was equipped a gradient microfluidic channel which enabled testing combination of medications. This is the first device where chemotherapyimmunotherapy medications can be tested together. This will improve the screening timeserve as an alternative for patients in therapeutic try.

The IAMA system with Automatic Virtual Metrology (AVM) technology can enable a 3D printing laser powder bed fusion (LPBF) machine to achieve the world＇s first innovative capability of online layer-to-layer quality predictionfeedback control, which improves the final product quality such as tensile strengthsurface roughnessrealize the goal of nearly Zero Defects of all products.

Due to the recent outbreak of COVID-19ongoing mutants, we developed a coronavirus variant protein microarray that includes all the current spike mutationsthe wild-type viral antigensapplied in detecting immune responsesdrug specificities. We filed the patientspublished our findings in the Analytical Chemistry. Besides massive production of the coronavirus protein microarray, we also collaborated with a scanner company, developed artificial intelligence software,collaborated with suppliers. We are welcome any sort of collaboration.

The Center has established a high biological conservation center for the breeding of P. monodon shrimp seeds that meet international standards,has integrated more than 40 professional technologies to select P. monodon shrimp strains that are specific pathogens free (SPF)have special traits for stocking in the three-hectare commercial P. monodon shrimp seed expansion test site in Hualien. This will provide a stable supply of good P. monodon shrimp seeds for use by the Taiwan aquaculture industry.

This technique combines deep learningdigital twin technologies to connect the designconstruction processto achieve construction-site rebar inspection, issue tracking, verificationmaintenance. This core component of this method is based on the image collection of rebar frames, digital twin creation, cooperative machine learning,the rebar inspection module development. The BIM model was automatically matchedcompared with the 4D digital twin with features of point cloudtime sequence.

首創(a)超薄型毫米波雙極化寬頻天線，較過去成果具厚度優勢(b)3D雙曲面空腔結構增加垂直貼片天線頻寬，較過去具頻寬優勢(c)毫米波多功能RFIC整合技術，大幅降低封測廠毫米波RFIC自動測試方案成本及(d)低變異量頻率可重置之寬頻切換式相移器，大幅切換頻帶時仍維持低振幅與相位變異量。

This technology proposes the preparation of Anodic Aluminum Oxide (AAO) templates on the surface of aluminum. The structural color of AAO is combined with lithography process to form patterns on the substrate, which can achieve high resolution on various aluminum substrate for decoration. Also, rewritable boardinvisible anti-counterfeiting technology are proposed for innovative applications by the low-cost, high-efficiency,high-saturation AAO structural colorpatterning.

Swelling at the extremities of the extremities is quite common in KD,it is one of the five major diagnostic criteria for KD. According to the results of our research, about 90 of the children will have this symptomis the most important symptoms of KD suggested by American Heart Association. Our research also found that albumin decreased,the degree of reduction is obviously positively related to the severity of the disease, so we have developed a theory that can detect water, hemethe optical method of oxygen. Provides a non-intrusive method for quick differentiation.

The Artificial Intelligence Underwater MonitoringManagement system is a system that integrates aquacultureinformation technology, which can collect image data of shrimp 24/7 by using IoT devices, then next enhance the images that collected under different light, water qualityturbidity. The system will recordanalyze the number, size, healtheating status of white shrimp, checking the amount of leftovers feed on the feeding tray, then manage the feeding strategycontrol water quality based on the above analysis.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infected human cell through ACE2. This study is developing a Pd nano-thin film electrode for electrochemical impedance spectroscopy. This invention can quantify neutralize antibody too. The detection platform can be used to screen 1 microliter anti-virus infection drug sample within 21 mins. Compare to SPR standard, our invention only needs 1/10000 concentration of ACE2reaction volume is ~1/30 than SPR. This invention can be adapted in detection of cancer markers,other emergency infection diseasestherapeutic drugs, too.

The present technology provides a method for diagnosing Alzheimer＇s disease, including the following steps: detecting a blood sample taken from a human subject suspected of having Alzheimer＇s disease,the biomarker level is selected from G72 protein level, SLC7A11 mRNA expression level,a combination of the two,compared with a preset standard, wherein if the detected biomarker level in the blood sample is higher than the preset standard, the Individual humans are thought to have Alzheimer＇s disease.

A portablemagnetic surface Raman-enhanced scattering (SERS) nanoparticle arrays platform would be developed for magnetic separationSERS detection of environmental, water,biomolecules samples. It is very simplerapid platform with portable Raman spectroscopy to detect samples (bacteria, virous, water toxicants) after magnetic separation processes, without labelling the fluorescence dye,the limit of detection concentration is lower than 10-8 M.

This study proposes a double elastic steel sheet vibration energy harvester system (DESS VEH), where the vibration generated by the deformation and/or clap of two magnetic excited elastic steel sheets is assisted by a piezo-patch to generate electric energy. The proposed model can be applied on an unmanned helicopterany other multi-rotor drones. The downstream flows from the rotors are collecteddriving a wind-turbine to excite the vibration steel sheet(s) systemgenerating electric power.

We propose a novel AI-based few-shot self-supervised learning method for automatic optical inspection image quality assessmentcomponent detection based on only few training images. Our method iteratively learns the feature representations of the components by using self-similarity of these components. With the large number of self-learned representations, the appearance variations of each component are then effectively learned in the AI model for component detectionmeasurement. The computation complexity of our method is significantly lower than that of deep learning methods.

We use AI analysissimulation technology to produce decision-making parameters according to requirements input to designdevelop different materialspatterning technologies. The technology can quickly produce world-leading biomimetic double-sided tapes with broad applications, even in harsh environments. In addition to assisting in advanced industrial processindustrial automation for companies, the usage of the double-sided tapes combined with bionic grippers becomes a system for lifting thin plate objectseasily adheringpeeling repeatedly.

This technology contains two important sub-technologies. The first is the development of ultra-high performance concrete composite materials, mastering the 3D printing related properties,developing printable concrete that can actually be used for printing. The second is the simulation, analysisoptimization technology of the concrete 3D printing process. The infinite element method developed by ourselves is used to simulateanalyze the concrete 3D printing process,the geometric of the object is considered to optimize the printing pathparameters.

We propose a semantic-guided framework (FashionOn+) that generates image-based virtual try-on results with arbitrary poses. FashionOn+ contains three stages: (I) conducts the semantic segmentation to have the prior knowledge of body parts for rendering the corresponding texture in stage (II). (III) refines two salient regions, i.e., faceclothes, to generate high-quality results. With the novel architecture, we win first place in the Multi-pose Virtual Try-on Challenge in CVPR, 2020. Further, we tackle the low-resolution limitation (256x192)achieve high-resolution results (640x480).

The proposed technologies apply deep learning methodsbig clinical data for lung cancer decision support. The system consists of: 1) CT radiogenomicspatho-genomics for automatic cancer detectionprediction of EGFR mutation 2) demographics to predict survival 3) genomics to predict cancer recurrencemetastasis,4) drug response inference for the best target therapy option. The modules are constructed on an AI-based Clinical Decision Support System (CDSS-SDM). The system is expected to provide physicianspatients with personalized medication recommendations.

We have developed an innovative acoustic vortex technology VorteSonic for treatment of VTE, with the first go-to-market product of a catheter-based acoustic vortex thrombolytic device. Acoustic vortex produces turbulent flow around the thrombus, resulting in effective thrombolysis by creating loose structure of thrombustherefore enhancing deep penetration of thrombolytic drugs into the thrombus. Advantages of the acoustic vortex technology is reduction of hemorrhagic riskimprovement of safetyprognosis. Feedback system can be integrated for intelligent therapy in the future.