基於聯合嵌入之雙手配對與追蹤系統;A Hands Pairing and Tracking System base on Associative Embedding

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/86426

Title:	基於聯合嵌入之雙手配對與追蹤系統;A Hands Pairing and Tracking System base on Associative Embedding
Authors:	沈桓慶;Shen, Huan-Ching
Contributors:	資訊工程學系
Keywords:	深度學習;偵測系統;手勢追蹤;物件偵測;類神經網路;聯合嵌入;deep learning;detection system;gesture tracking;object detection;neural network;associative embedding
Date:	2021-05-21
Issue Date:	2021-12-07 12:49:19 (UTC+8)
Publisher:	國立中央大學
Abstract:	手部追蹤旨在預測影像序列中多個手的軌跡，對於空中手寫、手語辨識及手勢辨識等應用具有重要的意義，而將雙手分組可以使上述應用實現更複雜的功能。本論文提出基於YOLOv3和聯合嵌入的方法，整合多目標追蹤和關節點檢測的單階段類神經網路模型和演算法，實現實時的多人雙手追蹤。;Hand tracking aims to predict the trajectory of multiple hands in an image sequence, which is of great significance for applications such as air handwriting, sign language recognition and gesture recognition, and grouping the hands can enable the above applications to achieve more complex functions. This paper proposes a single-stage neural network model and algorithm based on YOLOv3 and associative embedding, integrating multi-target tracking and joint point detection, to achieve real-time multi-person hand tracking.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	78	View/Open

社群 sharing