VIDEO LOGO DETECTION BY DEEP-TRANSFER ACTIVE LEARNING

HANG SU; GUOPING QIU

首页> 外文期刊>Discrete and continuous dynamical systems, Series S >VIDEO LOGO DETECTION BY DEEP-TRANSFER ACTIVE LEARNING

【24h】

VIDEO LOGO DETECTION BY DEEP-TRANSFER ACTIVE LEARNING

机译：VIDEO LOGO DETECTION BY DEEP-TRANSFER ACTIVE LEARNING

获取原文

获取原文并翻译 | 示例

掌桥外文数据库（机构版） >>

开具论文收录证明 >>

文献代查 >>

页面导航

摘要
著录项
相关主题

摘要

Brand logo detection is a special aspect of machine vision. However,Video logo detection benchmarks are scarce in the public domain. Weexploit the power of a deep convolutional neural network (DCNN) and leverageestablished datasets related to existing applications to develop a deep-transferactive-learning (DTAL) algorithm to select the most valuable samples so thatthe smallest number possible needs to be labeled to achieve maximum performanceimprovements for video object detection model training. By exploitingthe possible shared deep feature space between static and video datasetsthrough transfer learning based on highly adaptable DCNN features, DTAL implementsdiversity-based active learning to select the most informative samplesfrom a sequence of similar image frames for video object detection. We successfullyapply the new DTAL algorithm to implement active learning for logodetection from live streaming sports videos as well as pedestrian and face detectionfrom video data. We show that DTAL is a better active-learning methodthan state-of-the-art deep-learning-based active-learning object detection techniques.We also contribute one of the largest video-based logo resources, theSports Match Video Logo (SMVL) dataset, to facilitate general logo detectionresearch using transfer- and active-learning applications for video objectdetection.

著录项

来源
《Discrete and continuous dynamical systems, Series S》 |2023年第7期|1810-1826|共17页
作者
HANG SU; GUOPING QIU;
展开▼
作者单位

College of Electronic and Information Engineering, Shenzhen University, China;

College of Electronic and Information Engineering, Shenzhen University, China,Guangdong Key Lab for Intelligent Information Processing, China,Shenzhen Institute of AI and Robotics for Society, China,Pengcheng Laboratory, China,School of Computer Science,;

展开▼
收录信息
原文格式 PDF
正文语种英语
中图分类
关键词
Object detection; logo detection; video dataset; deep learning; active learning; transfer learning;

VIDEO LOGO DETECTION BY DEEP-TRANSFER ACTIVE LEARNING

摘要

著录项

相关主题

期刊订阅