Neurocle is a deep learning vision inspection software company aiming to be the 'Adobe of deep learning,' providing user-friendly solutions for non-experts [Database, 32]. The company integrates AI deep learning technology with computer vision to solve complex image interpretation problems across various industries. It specializes in auto deep learning vision inspection solutions that achieve superior inspection accuracy with minimal resources. Established in June 2019, Neurocle has grown into a global AI startup through continuous technological advancements.Neurocleisadeeplearningvisioninspectionsoftwarecompanyaimingtobethe'Adobeofdeeplearning,'providinguser-friendlysolutionsfornon-experts[Database,32].ThecompanyintegratesAIdeeplearningtechnologywithcomputervisiontosolvecompleximageinterpretationproblemsacrossvariousindustries.Itspecializesinautodeeplearningvisioninspectionsolutionsthatachievesuperiorinspectionaccuracywithminimalresources.EstablishedinJune2019,NeuroclehasgrownintoaglobalAIstartupthroughcontinuoustechnologicaladvancements.
Key Products/TechnologiesKeyProducts/Technologies
Neuro-T: A deep learning vision model training and creation software, enabling non-experts to easily generate high-performance models through an intuitive GUI and auto deep learning algorithms. It supports the entire process from image data preprocessing, labeling, model training and creation, to performance evaluation, with a Patch Classification model for detecting minute defects in high-resolution images. The solution also includes a Generative AI (GAN) model and Generation Center to address data scarcity by creating virtual defect images.Neuro-T:Adeeplearningvisionmodeltrainingandcreationsoftware,enablingnon-expertstoeasilygeneratehigh-performancemodelsthroughanintuitiveGUIandautodeeplearningalgorithms.Itsupportstheentireprocessfromimagedatapreprocessing,labeling,modeltrainingandcreation,toperformanceevaluation,withaPatchClassificationmodelfordetectingminutedefectsinhigh-resolutionimages.ThesolutionalsoincludesaGenerativeAI(GAN)modelandGenerationCentertoaddressdatascarcitybycreatingvirtualdefectimages.
Neuro-X: A vision software designed for deep learning experts, offering various hyperparameter tuning functionalities for flexible model performance optimization. It efficiently supports data management and model creation processes, aiming to standardize deep learning utilization capabilities.Neuro-X:Avisionsoftwaredesignedfordeeplearningexperts,offeringvarioushyperparametertuningfunctionalitiesforflexiblemodelperformanceoptimization.Itefficientlysupportsdatamanagementandmodelcreationprocesses,aimingtostandardizedeeplearningutilizationcapabilities.
Neuro-R: A real-time inference (Runtime) library or API that applies deep learning models created with Neuro-T to actual industrial sites. It is compatible with a wide range of devices and systems, from low-spec embedded boards to high-spec CPUs and GPUs, ensuring fast inspection speeds optimized for in-line processes. Compatibility with major programming languages like C++, Python, and C# is a key feature.Neuro-R:Areal-timeinference(Runtime)libraryorAPIthatappliesdeeplearningmodelscreatedwithNeuro-Ttoactualindustrialsites.Itiscompatiblewithawiderangeofdevicesandsystems,fromlow-specembeddedboardstohigh-specCPUsandGPUs,ensuringfastinspectionspeedsoptimizedforin-lineprocesses.CompatibilitywithmajorprogramminglanguageslikeC++,Python,andC#isakeyfeature.
Neuro-T Engine: A training-dedicated deep learning engine offering core Auto Deep Learning functionality without a GUI, utilized for building MLOps environments and embedding AI learning within industrial equipment.Neuro-TEngine:Atraining-dedicateddeeplearningengineofferingcoreAutoDeepLearningfunctionalitywithoutaGUI,utilizedforbuildingMLOpsenvironmentsandembeddingAIlearningwithinindustrialequipment.
Neuro-EDU: An AI education software designed to enhance the accessibility of deep learning technology and support its application in educational fields.Neuro-EDU:AnAIeducationsoftwaredesignedtoenhancetheaccessibilityofdeeplearningtechnologyandsupportitsapplicationineducationalfields.
Core AdvantagesCoreAdvantages
Provision of intuitive GUI-based no-code software, allowing non-experts to easily create and utilize deep learning vision models.ProvisionofintuitiveGUI-basedno-codesoftware,allowingnon-expertstoeasilycreateandutilizedeeplearningvisionmodels.
Proprietary Auto Deep Learning algorithm that automatically optimizes deep learning model structures and training parameters, enabling the creation of high-performance models without specialized deep learning knowledge.ProprietaryAutoDeepLearningalgorithmthatautomaticallyoptimizesdeeplearningmodelstructuresandtrainingparameters,enablingthecreationofhigh-performancemodelswithoutspecializeddeeplearningknowledge.
High inspection accuracy and performance, capable of precisely detecting minute defects in high-resolution images through advanced patch-based algorithm technology, overcoming limitations of traditional models.Highinspectionaccuracyandperformance,capableofpreciselydetectingminutedefectsinhigh-resolutionimagesthroughadvancedpatch-basedalgorithmtechnology,overcominglimitationsoftraditionalmodels.
Fast Retraining feature that minimizes retraining time while maintaining existing model performance when training data changes, enhancing model maintenance convenience.FastRetrainingfeaturethatminimizesretrainingtimewhilemaintainingexistingmodelperformancewhentrainingdatachanges,enhancingmodelmaintenanceconvenience.
Ability to solve data scarcity problems by utilizing Generative AI (GAN) models to create large quantities of realistic virtual defect data, thereby improving model training efficiency.AbilitytosolvedatascarcityproblemsbyutilizingGenerativeAI(GAN)modelstocreatelargequantitiesofrealisticvirtualdefectdata,therebyimprovingmodeltrainingefficiency.
AI-based automatic labeling functionality that significantly reduces labeling resources required for deep learning model creation and maximizes labeling consistency.AI-basedautomaticlabelingfunctionalitythatsignificantlyreduceslabelingresourcesrequiredfordeeplearningmodelcreationandmaximizeslabelingconsistency.
Flexible compatibility with various hardware platforms including embedded boards, CPUs, GPUs, and NPUs, as well as multiple programming languages like C++, Python, and C#, supporting a wide range of application environments.Flexiblecompatibilitywithvarioushardwareplatformsincludingembeddedboards,CPUs,GPUs,andNPUs,aswellasmultipleprogramminglanguageslikeC++,Python,andC#,supportingawiderangeofapplicationenvironments.
Focus on standard software tools that eliminate the need for customization, allowing customers to easily create and apply AI models tailored to their needs, thereby securing market competitiveness through versatility.Focusonstandardsoftwaretoolsthateliminatetheneedforcustomization,allowingcustomerstoeasilycreateandapplyAImodelstailoredtotheirneeds,therebysecuringmarketcompetitivenessthroughversatility.
Target IndustrieTargetIndustrie
Manufacturing (including secondary batteries, semiconductors, displays, automotive, tires, precision electronic components, heavy industry & metal, food & beverage)Manufacturing(includingsecondarybatteries,semiconductors,displays,automotive,tires,precisionelectroniccomponents,heavyindustry&metal,food&beverage)
Medical (including university hospitals, general hospitals, CT, MRI, X-Ray, ultrasound, endoscope, PET-CT, electron microscope)Medical(includinguniversityhospitals,generalhospitals,CT,MRI,X-Ray,ultrasound,endoscope,PET-CT,electronmicroscope)
LogisticsLogistics
Security (CCTV)Security(CCTV)
EducationEducation
Major MarketsMajorMarkets
Japan, China, Taiwan, Malaysia, Vietnam, Singapore, ThailandJapan,China,Taiwan,Malaysia,Vietnam,Singapore,Thailand
Registration of key technology patents related to Auto Deep Learning algorithms.RegistrationofkeytechnologypatentsrelatedtoAutoDeepLearningalgorithms.
Recipient of the 'Innovators Awards' by Vision Systems Design for 5 consecutive years (including Silver in 2021, Bronze in 2022, and Platinum in 2023).Recipientofthe'InnovatorsAwards'byVisionSystemsDesignfor5consecutiveyears(includingSilverin2021,Bronzein2022,andPlatinumin2023).
Awarded the 'ImpaCT-ech Grand Prize' by the Ministry of Science and ICT.Awardedthe'ImpaCT-echGrandPrize'bytheMinistryofScienceandICT.
Selected as a 'SW Preliminary High-Growth Company' for two consecutive years by the Ministry of Science and ICT.Selectedasa'SWPreliminaryHigh-GrowthCompany'fortwoconsecutiveyearsbytheMinistryofScienceandICT.
Sole Asian company to be a Top 10 Finalist in 'The Cowen Startup Challenge'.SoleAsiancompanytobeaTop10Finalistin'TheCowenStartupChallenge'.
Recipient of the 'AI Korea Grand Prize' (Minister of SMEs and Startups Award).Recipientofthe'AIKoreaGrandPrize'(MinisterofSMEsandStartupsAward).
Awarded the 'Challenge K-Startup Excellence Award' (Minister of Culture, Sports and Tourism Award).Awardedthe'ChallengeK-StartupExcellenceAward'(MinisterofCulture,SportsandTourismAward).
Named a 'Gartner Cool Vendor in AI for Computer Vision'.Nameda'GartnerCoolVendorinAIforComputerVision'.
Recipient of the 'Korea Challenge Participation Award' at 'Try Everything 2022'.Recipientofthe'KoreaChallengeParticipationAward'at'TryEverything2022'.
Selected as '2023 Emerging AI+X Top 100' by the Artificial Intelligence Industry Association.Selectedas'2023EmergingAI+XTop100'bytheArtificialIntelligenceIndustryAssociation.
Awarded the 'AI Technology Award' (Minister of Science and ICT Award).Awardedthe'AITechnologyAward'(MinisterofScienceandICTAward).
Certified as a Venture Company.CertifiedasaVentureCompany.
Introduction
Location
South Korea, Seoul, Gangdong-gu, 고덕 비즈밸리로 4길 35 디엠스퀘어
클릭하여 위치 살펴보기
Information
South Korea, Seoul, Gangdong-gu, 고덕 비즈밸리로 4길 35 디엠스퀘어