Spatio-temporal Data Augmentation for Visual Surveillance
نویسندگان
چکیده
Visual surveillance aims to stably detect a foreground object using continuous image acquired from fixed camera. Recent deep learning methods based on supervised show superior performance compared classical background subtraction algorithms. However, there is still room for improvement in static foreground, dynamic background, hard shadow, illumination changes, camouflage, etc. In addition, most of the learning-based operates well environments similar training. If testing are different training ones, their degrades. As result, additional those operating required ensure good performance. Our previous work which uses spatio-temporal input data consisted number past images, images and current showed promising results training, although it simple U-NET structure. this paper, we propose augmentation technique suitable visual same network used our work. learning, techniques deal with spatial-level use classification detection. new method dimension Two adjusting model proposed. Through this, shown that can be improved difficult areas such as ghost objects, studies. quantitative qualitative evaluation SBI, LASIESTA, own dataset, gives algorithms
منابع مشابه
Spatio-Temporal Disease Surveillance
There are two principal impediments in statistical process control methods for the detection of bio-terrorism events: firstly, these methods aggregate over space by examining total counts and thus ignore the spatial dimension of the task and secondly they fail to adjust for the usual (seasonal) behaviour of diseases (e.g., Steiner et al., 2011 where the focus is early detection of the start of ...
متن کاملSpatio-temporal Background Models for Outdoor Surveillance
Video surveillance in outdoor areas is hampered by consistent background motion which defeats systems that use motion to identify intruders. While algorithms exist for masking out regions with motion, a better approach is to develop a statistical model of the typical dynamic video appearance. This allows the detection of potential intruders even in front of trees and grass waving in the wind, w...
متن کاملSpatio-Temporal Visual Ontology
In order to semantically describe and structure the visual content of images and video scenes captured by cameras with arbitrary resolution and unknown calibration properties, we propose a spatio-temporal visual ontology (STVO). In Artificial Intelligence (AI), an ontology is an explicit specification of a conceptualization [4] and thus consists of a set of semantic concepts as well as the rela...
متن کاملContext-aware Modeling for Spatio-temporal Data Transmitted from a Wireless Body Sensor Network
Context-aware systems must be interoperable and work across different platforms at any time and in any place. Context data collected from wireless body area networks (WBAN) may be heterogeneous and imperfect, which makes their design and implementation difficult. In this research, we introduce a model which takes the dynamic nature of a context-aware system into consideration. This model is con...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Access
سال: 2021
ISSN: ['2169-3536']
DOI: https://doi.org/10.1109/access.2021.3135505