公开(公告)号：US20240193809A1

公开(公告)日：2024-06-13

申请号：US18287518

申请日：2022-05-09

Applicant: Northeastern University

Inventor： Sarah OSTADABBAS ,  Xiaofei HUANG ,  Nihang FU ,  Shuangjun LIU

IPC: G06T7/73 ,  G06T15/04

CPC classification number: G06T7/74 ,  G06T7/75 ,  G06T15/04 ,  G06T2207/10016 ,  G06T2207/10024 ,  G06T2207/10028 ,  G06T2207/10048 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/30196

Abstract: Methods are provided for estimating a pose of an infant using image analysis and artificial intelligence. A classifier is trained using a dataset containing hybrid synthetic and real infant pose data. Multi-stage invariant representation machine learning strategies are employed that transfer knowledge from adjacent domains of adult poses and synthetic infant images into a fine-tuned domain-adapted infant pose estimation model.