Samsung / SamsungAutomationStudio

Samsung Automation Studio is to provide development tools and execution environment that can easily configure application logic by connecting both Samsung service and 3rd party service. This project is to share the node for open source NodeRED developed by Samsung Automation Studio team to the community. If you are using nodered, you can easily install the node we provide. And you can use Samsung's IoT and AI-related services more easily, and you can have an extended experience in conjunction with your own services.
Apache License 2.0
56 stars 29 forks source link

Request to add vision AI nodes #82

Closed chickennight closed 9 months ago

chickennight commented 9 months ago

Hello. This is the 'Good I-Deer' team from the Samsung SW Academy For Youth 9th.

In the Samsung Automation Studio for Mashup (hereinafter referred to as 'SASM'), we provide object detection, face detection, face vectorization, vector comparing and vector database nodes that work with SmartThings devices.

The nodes we developed are as follows.

Node Name Description
Good Obejct Detection Extract object information from image
Good Webcam Object Detection Extract object information from Webcam
Good Face Detection Extract face information from image
Good Face Vectorization Extract feature points from face image and convert them into 512 dimension vectors
Good Calculate Consine Calculate cosine similarity between vectors
Good ChromaDB Use ChromaDB to store and query vector values

Special Thanks to Samsung Electronics mentor Sehwan Chang who led us during the SSDC project.


안녕하세요. 삼성 청년 SW 아카데미 9기의 'Good I-Deer' 팀 입니다.

Samsung Automation Studio for Mashup(이하 SASM)에서 SmartThings 기기와 연동하여 객체 인식, 얼굴 인식, 얼굴 벡터화, 벡터 값 비교 및 저장하는 노드의 개발을 진행했습니다.

저희가 개발한 노드는 다음과 같습니다.

노드 이름 기능
Good Obejct Detection 사진에서 객체 정보를 추출
Good Webcam Object Detection 웹캠에서 객체 정보를 추출
Good Face Detection 사진에서 얼굴 정보를 추출
Good Face Vectorization 얼굴 사진에서 특징점을 추출하여 512개의 벡터로 변환
Good Calculate Consine 두 벡터 사이의 코사인 유사도를 계산
Good ChromaDB ChromaDB를 이용해서 벡터 값을 저장 및 검색

SSDC 프로젝트 기간동안 저희를 이끌어주신 삼성전자 장세환 멘토님께 감사드립니다.