AI-Based Projects
인공지능 기술을 활용한 혁신적인 프로젝트들을 만나보세요. 머신러닝부터 딥러닝까지 다양한 AI 솔루션을 제공합니다.
K-Bank Unstructured Data-Based Credit Risk Assessment and Fraud Detection AI Model
Are you losing countless potential customers who could become 'future VIP clients' to competitors simply because they lack financial records? Existing systems are helpless against evolving intelligent fraud tactics, leading to declining customer trust and direct annual losses of billions of won.
Personalized Marketing Video Template Recommendation Service
Marketers spent significant time finding the optimal template matching campaign goals and target customers among numerous video templates. This created bottlenecks in video production, and using generic templates led to low customer engagement and conversion rates.
AI Microscope Image Analysis-Based Cell Reactivity Measurement
In the early stages of drug development, researchers' manual analysis of electron microscope images required considerable time and effort. Additionally, analysis results varied depending on researchers' subjectivity, making consistency difficult to ensure, and subtle cellular morphological changes were often missed, making it hard to accurately determine drug reactivity.
Immigrant Women Translation Expert Matching Curation Algorithm
Immigrant women with professional translation capabilities struggled to find stable jobs due to career interruptions. Simultaneously, companies were inefficiently spending time and money finding optimal translators who understand not only language proficiency but also specific industry expertise and target market cultural backgrounds.
Video-Based Music Recommendation and Signal Music Generation AI
Finding suitable background music for video content consumed significant time, and it was difficult to find music that satisfies both brand identity and video atmosphere in vast libraries. Additionally, creating brand-specific signal sounds required high costs and specialized personnel, presenting high entry barriers.
AI-Based Audio Source Separation and Dolby Atmos Spatial Audio Synthesis
Remixing existing master audio sources into spatial audio like Dolby Atmos was virtually impossible without original multitrack (stem) files. Traditional instrument separation technology without AI had severe sound quality degradation and artifact issues, and manual work required enormous time and costs. This created significant barriers to converting vast existing audio libraries into high-quality spatial audio.
Foresee: Artist-Label Matching and Growth Strategy AI
Emerging artists struggled to gain label attention in the vast market and lacked clear data-driven strategies for career development (e.g., which countries to target, which genres to focus on). Labels also had to rely heavily on resources and intuition to discover potential artists and predict success probability.
Real Gen: Natural Language-Based LED Wall VFX Background Generation AI
While virtual production using massive LED Walls reduced filming time compared to traditional green screens, creating background assets still required enormous time and costs. Especially when director requirements changed on set (e.g., 'make that mountain higher', 'darken the sky'), immediate modifications were impossible, delaying production or requiring reliance on post-production.
AI-Based Surgical Medical Tool Recognition and Training System
Junior nurses suffered from pressure to quickly identify and prepare hundreds of complex surgical instruments. This led to mistakes during handover processes or communication errors with senior nurses, increasing emotional stress and reducing surgical preparation efficiency.
Besides: Disabled Person Resume Automation and Recruitment Brokerage AI
Disabled job seekers struggled greatly to effectively express their strengths and potential in standardized resume formats. Additionally, companies had limitations in discovering suitable talent focused on job competencies and aptitudes rather than disability types, leading to employment inequality issues for people with disabilities.