Video Segmentation using Keywords

机译：使用关键字的视频细分

获取原文

页面导航

摘要
著录项
相似文献
相关主题

摘要

At DAVIS-2016 Challenge, many state-of-art video segmentation methods achieve potential results, but they still much depend on annotated frames to distinguish between background and foreground. It takes a lot of time and efforts to create these frames exactly. In this paper, we introduce a method to segment objects from video based on keywords given by user. First, we use a real-time object detection system - YOLOv2 to identify regions containing objects that have labels match with the given keywords in the first frame. Then, for each region identified from the previous step, we use Pyramid Scene Parsing Network to assign each pixel as foreground or background. These frames can be used as input frames for Object Flow algorithm to perform segmentation on entire video. We conduct experiments on a subset of DAVIS-2016 dataset in half the size of its original size, which shows that our method can handle many popular classes in PASCAL VOC 2012 dataset with acceptable accuracy, about 75.03%. We suggest widely testing by combining other methods to improve this result in the future.

机译：在DAVIS-2016挑战赛上，许多先进的视频分割方法均取得了潜在的效果，但是它们仍然很大程度上依赖于带注释的帧来区分背景和前景。准确地创建这些框架需要花费大量时间和精力。在本文中，我们介绍了一种根据用户指定的关键字从视频中分割对象的方法。首先，我们使用实时对象检测系统YOLOv2来识别包含对象的区域，这些对象的标签与第一帧中的给定关键字相匹配。然后，对于从上一步确定的每个区域，我们使用金字塔场景解析网络将每个像素分配为前景或背景。这些帧可用作对象流算法的输入帧，以对整个视频执行分割。我们对DAVIS-2016数据集的子集进行了实验，其大小仅为原始大小的一半，这表明我们的方法可以以可接受的准确度处理PASCAL VOC 2012数据集中的许多流行类，约为75.03％。我们建议通过结合其他方法来进行广泛的测试，以在将来改善此结果。

著录项

来源
《International conference on machine vision》|2017年|106960U.1-106960U.8|共8页
会议地点
作者
Vinh Ton-That; Chi-Tai Vong; Xuan-Truong Nguyen-Dao; Minh-Triet Tran;
展开▼
作者单位

展开▼
会议组织
原文格式 PDF
正文语种
中图分类
关键词
video segmentation; object detection; object flow; saliency;

机译：视频分割;目标检测对象流;显着性;

相似文献

外文文献
中文文献
专利

1. Assisted keyword indexing for lecture videos using unsupervised keyword spotting [J] . Kanadje Manish, Miller Zachary, Agarwal Anurag, Pattern recognition letters . 2016,第Feba1期

机译：使用无监督关键字发现功能对演讲视频进行辅助关键字索引
2. Classification of educational videos by using a semi-supervised learning method on transcripts and keywords [J] . Stoica Alexandru Stefan, Heras Stella, Palanca Javier, Neurocomputing . 2021,第Octa7期

机译：通过在成绩单和关键字上使用半监督学习方法进行教育视频的分类
3. Keyword-Propagation-Based Information Enriching and Noise Removal for Web News Videos [J] . Jun Zhang, Xiaoming Fan, Jianyong Wang, SIGKDD explorations . 2012,第CDaROM期

机译：基于关键词传播的网络新闻视频信息丰富与消噪
4. Video Segmentation using Keywords [C] . Vinh Ton-That, Chi-Tai Vong, Xuan-Truong Nguyen-Dao, International Conference on Machine Vision . 2018

机译：使用关键字进行视频分割
5. Video partitioning and video segmentation in content-based video databases. [D] . Xu, Haifeng. 2001

机译：基于内容的视频数据库中的视频分区和视频分段。
6. Real-Time Instance Segmentation of Traffic Videos for Embedded Devices [O] . Ruben Panero Martinez, Ionut Schiopu, Bruno Cornelis, 2021

机译：嵌入式设备交通视频的实时实例分割
7. Occupational therapy intervention with a child is based upon an understanding and appreciation of normal development. Knowledge of current concepts and theories related to child development is essential when occupational therapist evaluates children. This background information helps therapist to plan intervention for the child. The aim of this study is to make observation video about development of about one year old child. The purpose of my study is to help occupational therapy students learn about child development. My study is practice-based thesis. It includes product, which is the observation video and study rapport. I describe my whole process in my rapport. The process includes different kinds of stages. First, I studied those theories of child development, which are used in the studies of occupational therapy for children. These theories are Moseys Developmental Frame of Reference and the theory of development according to Sensory Integration Theory. These theories are the frames of reference of my study. I organize the child development areas according to child occupations and skills. Then I start to plan, film and edit my video based on the theories of child development and the principles of making a video. In my rapport I describe all the stages of my study and explain the sequence and the content of the stages. I also evaluate the process of my study. In the observation video you can see those stages of development where about one year old child is based on the frames of reference, which I have used in my study. I believe that my observation video can at least be good for inspiring occupational therapy students learning about child development. Keywords child development, learning, observation video [O] . Lehtinen Ann-Mari 2006

机译：对儿童的职业治疗干预基于对正常发育的理解和欣赏。当职业治疗师评估儿童时，与儿童发育相关的当前概念和理论的知识必不可少。这些背景信息可帮助治疗师为孩子计划干预措施。这项研究的目的是制作有关约一岁儿童发育的观察视频。我研究的目的是帮助职业治疗学生学习儿童成长。我的研究是基于实践的论文。它包括产品，这是观察视频和学习融洽的关系。我以融洽的方式描述我的整个过程。该过程包括不同阶段。首先，我研究了有关儿童发育的理论，这些理论被用于儿童的职业治疗研究中。这些理论是Moseys发展参考框架和根据感觉统合理论的发展理论。这些理论是我研究的参考框架。我根据儿童职业和技能组织儿童发展领域。然后，我根据儿童发育理论和视频制作原理开始计划，拍摄和编辑视频。在融洽的关系中，我描述了学习的所有阶段，并解释了这些阶段的顺序和内容。我还评估了我的学习过程。在观察视频中，您可以看到那些发展阶段，其中大约一岁的孩子基于我的研究框架。我相信，我的观察视频至少可以对激发职业治疗的学生学习儿童发育有帮助。关键字儿童发展，学习，观察视频

Video Segmentation using Keywords

摘要

著录项

相似文献

相关主题

期刊订阅