Ponce AI

Ponce AI

科技、信息和网络

San Francisco,CA 1,965 位关注者

Smarter medical consultations

关于我们

Ponce.ai lets aesthetic medical providers provide objective data during their consultations to validate their claims and increase treatment sales.

网站
https://ponce.ai
所属行业
科技、信息和网络
规模
2-10 人
总部
San Francisco,CA
类型
私人持股
创立
2022

地点

Ponce AI员工

动态

  • 查看Ponce AI的公司主页,图片

    1,965 位关注者

    We're excited to continue advancing healthcare experiences using the latest advances in AI and CV.

    查看Saniya Lande的档案,图片

    Data Analyst | Python, SQL, Tableau, Power BI, Machine Learning, Data Analysis, LLMs

    As I conclude my internship at Ponce AI, I wanted to share some insights into the fascinating projects I’ve had the privilege to work on, especially the Lip Type Clustering initiative, which was both challenging and immensely rewarding. Here's a glimpse into this innovative endeavor: ? - Data Collection: Embraced the FFHQ dataset along with diverse user-uploaded images to ensure a broad spectrum of lip types. - Pre-Processing: Used Mediapipe for precise lip region detection and cropping, focusing on the area crucial for accurate classification. - Efficient Clustering: Implemented Annoy Index for organizing lip images into clusters based on similarity, which was pivotal for categorizing different lip structures. - Quality Evaluation and Insight: Initially used Dunn Index and Silhouette Score, then transitioned to a more direct measurement of cluster relevance based on centroid distance. - Outcome: Successfully categorized lips into distinct clusters, significantly enhancing the realism of cosmetic simulations. - Operational Workflow: This clustering system can be integrated within the image generation pipeline, allowing for any input image to be automatically assigned to a cluster. This assignment can then inform the subsequent simulation processes, ensuring that each simulation was tailored to the specific characteristics of the input image. Other Noteworthy Projects: - Analyzed input image quality to differentiate between sharp and blurry images, ensuring high-quality inputs for the stable diffusion pipeline. - Developed an AI-powered RAG-based chatbot for medical consultations, improving customer interactions and query handling. I am immensely grateful to sugam arora, Rudolph Layko, and the entire team at Ponce AI for their guidance and mentorship. This experience has been a significant milestone in my professional journey, and I am excited to carry these learnings forward in my career. #DataScience #MachineLearning #ComputerVision #imageprocessing #InternshipExperience #InnovationInTech #GenerativeAI #stablediffusion #largelanguagemodels #chatbot

相似主页