scholarly journals Saliency Detection with Moving Camera via Background Model Completion

Sensors ◽  
2021 ◽  
Vol 21 (24) ◽  
pp. 8374
Author(s):  
Yupei Zhang ◽  
Kwok-Leung Chan
Keyword(s):  
Deep Learning ◽  
Background Subtraction ◽  
Moving Objects ◽  
Saliency Detection ◽  
False Negative ◽  
Background Model ◽  
Dynamic Background ◽  
Video Completion ◽  
Model Completion ◽  
Background Image

Detecting saliency in videos is a fundamental step in many computer vision systems. Saliency is the significant target(s) in the video. The object of interest is further analyzed for high-level applications. The segregation of saliency and the background can be made if they exhibit different visual cues. Therefore, saliency detection is often formulated as background subtraction. However, saliency detection is challenging. For instance, dynamic background can result in false positive errors. In another scenario, camouflage will result in false negative errors. With moving cameras, the captured scenes are even more complicated to handle. We propose a new framework, called saliency detection via background model completion (SD-BMC), that comprises a background modeler and a deep learning background/foreground segmentation network. The background modeler generates an initial clean background image from a short image sequence. Based on the idea of video completion, a good background frame can be synthesized with the co-existence of changing background and moving objects. We adopt the background/foreground segmenter, which was pre-trained with a specific video dataset. It can also detect saliency in unseen videos. The background modeler can adjust the background image dynamically when the background/foreground segmenter output deteriorates during processing a long video. To the best of our knowledge, our framework is the first one to adopt video completion for background modeling and saliency detection in videos captured by moving cameras. The F-measure results, obtained from the pan-tilt-zoom (PTZ) videos, show that our proposed framework outperforms some deep learning-based background subtraction models by 11% or more. With more challenging videos, our framework also outperforms many high-ranking background subtraction methods by more than 3%.

Algorithm of Color Detection for Moving Video Objects Based on Mode Matching

2012 ◽  
Vol 239-240 ◽  
pp. 1000-1003
Author(s):  
Zhao Quan Cai ◽  
Hui Hu ◽  
Tao Xu ◽  
Wei Luo ◽  
Yi Cheng He
Keyword(s):  
Background Subtraction ◽  
Moving Objects ◽  
Color Matching ◽  
Human Beings ◽  
Current Image ◽  
Color Recognition ◽  
Recognition Ability ◽  
Color Identification ◽  
The Difference ◽  
Background Image

It is urgent to study how to effectively identify color of moving objects from the video in the information era. In this paper, we present the color identification methods for moving objects on fixed camera. One kind of the methods is background subtraction that recognizes the foreground objects by compare the difference of pixel luminance between the current image and the background image at the same coordinates. Another kind is based on the statistics of HSV color and color matching which makes the detection more similar to the color identification of the human beings. According to the experiment results, after the completion of the background modelling, our algorithm of background subtraction, statistics of the HSV color and the color matching have strong color recognition ability on the moving objects of video.

scholarly journals Moving Object Detection for Dynamic Background Scenes Based on Spatiotemporal Model

2017 ◽  
Vol 2017 ◽  
pp. 1-9 ◽  
Author(s):  
Yizhong Yang ◽  
Qiang Zhang ◽  
Pengfei Wang ◽  
Xionglou Hu ◽  
Nengju Wu
Keyword(s):  
Object Detection ◽  
Moving Objects ◽  
Moving Object Detection ◽  
Moving Object ◽  
Background Model ◽  
Dynamic Background ◽  
Spatiotemporal Model ◽  
Temporal Features ◽  
Common Technique ◽  
Public Datasets

Moving object detection in video streams is the first step of many computer vision applications. Background modeling and subtraction for moving detection is the most common technique for detecting, while how to detect moving objects correctly is still a challenge. Some methods initialize the background model at each pixel in the first N frames. However, it cannot perform well in dynamic background scenes since the background model only contains temporal features. Herein, a novel pixelwise and nonparametric moving object detection method is proposed, which contains both spatial and temporal features. The proposed method can accurately detect the dynamic background. Additionally, several new mechanisms are also proposed to maintain and update the background model. The experimental results based on image sequences in public datasets show that the proposed method provides the robustness and effectiveness in dynamic background scenes compared with the existing methods.

scholarly journals Real-bogus classification for the Zwicky Transient Facility using deep learning

2019 ◽  
Vol 489 (3) ◽  
pp. 3582-3590 ◽  
Author(s):  
Dmitry A Duev ◽  
Ashish Mahabal ◽  
Frank J Masci ◽  
Matthew J Graham ◽  
Ben Rusholme ◽  
...  
Keyword(s):  
Deep Learning ◽  
False Positive ◽  
Large Scale ◽  
Moving Objects ◽  
False Negative ◽  
Efficient Manner ◽  
Astronomical Surveys ◽  
Comparable Performance ◽  
Cost Efficient ◽  
Initial Results

ABSTRACT Efficient automated detection of flux-transient, re-occurring flux-variable, and moving objects is increasingly important for large-scale astronomical surveys. We present braai, a convolutional-neural-network, deep-learning real/bogus classifier designed to separate genuine astrophysical events and objects from false positive, or bogus, detections in the data of the Zwicky Transient Facility (ZTF), a new robotic time-domain survey currently in operation at the Palomar Observatory in California, USA. Braai demonstrates a state-of-the-art performance as quantified by its low false negative and false positive rates. We describe the open-source software tools used internally at Caltech to archive and access ZTF’s alerts and light curves (kowalski ), and to label the data (zwickyverse). We also report the initial results of the classifier deployment on the Edge Tensor Processing Units that show comparable performance in terms of accuracy, but in a much more (cost-) efficient manner, which has significant implications for current and future surveys.

Block Background Subtraction Method Based on ViBe

2014 ◽  
Vol 556-562 ◽  
pp. 3549-3552
Author(s):  
Lian Fen Huang ◽  
Qing Yue Chen ◽  
Jin Feng Lin ◽  
He Zhi Lin
Keyword(s):  
Spatial Correlation ◽  
Background Subtraction ◽  
Video Sequence ◽  
Moving Objects ◽  
Moving Object ◽  
Background Model ◽  
Subtraction Method ◽  
Current Frame ◽  
The Difference ◽  
Set Up

The key of background subtraction which is widely used in moving object detecting is to set up and update the background model. This paper presents a block background subtraction method based on ViBe, using the spatial correlation and time continuity of the video sequence. Set up the video sequence background model firstly. Then, update the background model through block processing. Finally employ the difference between the current frame and background model to extract moving objects.

A New Moving Human Detection Method in Color Video Image

2012 ◽  
Vol 229-231 ◽  
pp. 1166-1170
Author(s):  
Tia Nai Wu ◽  
Yun Rong Wu ◽  
Yun Yu Wu
Keyword(s):  
Background Subtraction ◽  
Detection Method ◽  
Main Idea ◽  
Color Difference ◽  
Human Detection ◽  
Background Model ◽  
Video Frame ◽  
Difference Model ◽  
Background Updating ◽  
Background Image

Moving object detection is the basic of video applications such as computer vision, object recognition and tracking, surveillance security etc. Background subtraction and symmetrical differencing are the popular methods of motion detection. The main idea of them is to compare the current video frame with a specified background image or a background model or the next video frame. For background subtraction, the obtaining of initialization is crucial and many methods have been employed, so it is necessary to model background to adapt the changes of background. In this paper, the single gaussian modeling as the initialization background model combined with an improved linear alternate background updating method is proposed. And then, a novel moving human detection method which employs background subtraction and symmetrical differencing based on rgb color difference model is presented. The experimental results show that the detection method can detect moving human effectively and real-time.

scholarly journals Segmentation of Moving Objects using Numerous Background Subtraction Methods for Surveillance Applications

2020 ◽  
Vol 9 (3) ◽  
pp. 2553-2563
Keyword(s):  
Background Subtraction ◽  
Moving Objects ◽  
Tracking System ◽  
Real Life ◽  
Ground Truth ◽  
Computational Time ◽  
Dynamic Background ◽  
Illumination Changes ◽  
Benchmark Datasets ◽  
Surveillance Applications

Background subtraction is a key part to detect moving objects from the video in computer vision field. It is used to subtract reference frame to every new frame of video scenes. There are wide varieties of background subtraction techniques available in literature to solve real life applications like crowd analysis, human activity tracking system, traffic analysis and many more. Moreover, there were not enough benchmark datasets available which can solve all the challenges of subtraction techniques for object detection. Thus challenges were found in terms of dynamic background, illumination changes, shadow appearance, occlusion and object speed. In this perspective, we have tried to provide exhaustive literature survey on background subtraction techniques for video surveillance applications to solve these challenges in real situations. Additionally, we have surveyed eight benchmark video datasets here namely Wallflower, BMC, PET, IBM, CAVIAR, CD.Net, SABS and RGB-D along with their available ground truth. This study evaluates the performance of five background subtraction methods using performance parameters such as specificity, sensitivity, FNR, PWC and F-Score in order to identify an accurate and efficient method for detecting moving objects in less computational time.

scholarly journals Foreground algorithms for detection and extraction of an object in multimedia

2020 ◽  
Vol 10 (2) ◽  
pp. 1849
Author(s):  
Rekha V. ◽  
Natarajan K. ◽  
Innila Rose J.
Keyword(s):  
Background Subtraction ◽  
Feature Detection ◽  
Moving Objects ◽  
Traffic Monitoring ◽  
Mobile Platform ◽  
Current Frame ◽  
Conference Room ◽  
Detection Algorithms ◽  
Time Motion ◽  
Background Image

Background Subtraction of a foreground object in multimedia is one of the major preprocessing steps involved in many vision-based applications. The main logic for detecting moving objects from the video is difference of the current frame and a reference frame which is called “background image” and this method is known as frame differencing method. Background Subtraction is widely used for real-time motion gesture recognition to be used in gesture enabled items like vehicles or automated gadgets. It is also used in content-based video coding, traffic monitoring, object tracking, digital forensics and human-computer interaction. Now-a-days due to advent in technology it is noticed that most of the conferences, meetings and interviews are done on video calls. It’s quite obvious that a conference room like atmosphere is not always readily available at any point of time. To eradicate this issue, an efficient algorithm for foreground extraction in a multimedia on video calls is very much needed. This paper is not to just build Background Subtraction application for Mobile Platform but to optimize the existing OpenCV algorithm to work on limited resources on mobile platform without reducing the performance. In this paper, comparison of various foreground detection, extraction and feature detection algorithms are done on mobile platform using OpenCV. The set of experiments were conducted to appraise the efficiency of each algorithm over the other. The overall performances of these algorithms were compared on the basis of execution time, resolution and resources required.

scholarly journals Penerapan Metode Background Subtraction dengan Menggunakan Kandidat Sampling Background untuk Deteksi Kemacetan

2018 ◽  
Vol 5 (6) ◽  
pp. 731
Author(s):  
Cipto Prabowo ◽  
Zurnawita Zurnawita
Keyword(s):  
Information Technology ◽  
Background Subtraction ◽  
Traffic Congestion ◽  
Background Model ◽  
Raspberry Pi ◽  
Model Research ◽  
Background Traffic ◽  
Web Camera ◽  
Model Determination ◽  
Background Image

<p class="Abstrak"><span lang="IN">Penggunaan teknologi informasi untuk mengatasi kemacetan lalu lintas sudah banyak digunakan, salah satunya adalah dengan menggunakan CCTV. Citra dari CCTV diurai antara <em>foreground</em> (citra kendaraan, sebagai salah satu komponen penyebab kemacetan) dan <em>background</em>nya (citra latar lalu lintas), fokus penelitian salah satunya adalah cara menentukan <em>background model</em> dengan berbagai macam teknik. Pada penelitian ini mencoba memberikan alternatif penentuan <em>background model</em> dengan menerapkan kandidat <em>sampling background</em> sebagai <em>background model. </em>Lokasi penelitan pada simpang bypass ketaping padang dan pengambilan gambar menggunakan raspberry pi dan web camera dengan arah tembakan menyamping dan durasi tembakan tiap menit. Hasil yang didapatkan cukup memuaskan terutama pada kondisi pagi dan siang hari.</span></p><p class="Abstrak"> </p><p class="Abstrak"><strong><em>Abstract</em></strong></p><p class="Abstrak"> </p><p class="Abstrak"><em><span lang="IN">The use of information technology to overcome traffic congestion has been widely used, one of them is by using CCTV. The image of CCTV is parsed between the foreground (vehicle image, as one of the components of the congestion cause) and the background (traffic background image), the focus of research is one way of determining the background model with various techniques. In this research try to give alternative background model determination by applying background sampling candidate as background model. Research location at intersection bypass ketaping and taking pictures using raspberry pi and web camera with sideways shot direction and duration of shots every minute. The results obtained are quite satisfactory, especially in the condition of morning and afternoon.</span></em></p>

scholarly journals Method of Moving Region Detection for Static Camera

2018 ◽  
Vol 12 (6) ◽  
pp. 3626-3633
Author(s):  
Pravesh Kumar Goel ◽  
Paresh P. Kotak ◽  
Amit Gupta
Keyword(s):  
Computer Vision ◽  
Object Detection ◽  
Background Subtraction ◽  
Video Sequence ◽  
Moving Objects ◽  
Moving Object Detection ◽  
Region Detection ◽  
Computer Vision Applications ◽  
Object Center ◽  
Background Image

The moving object detection from a stationary video sequence is a primary task in various computer vision applications. In this proposed system; three processing levels are suppose to perform: detects moving objects region from the background image; reduce noise from the pixels of detected region and extract meaningful objects and their features (area of object, center point of area etc.). In this paper; background subtraction techniques is used for segments moving objects from the background image, which is capable for pixel level processing. Morphology operation (Erosion and dilation) are used to remove pixel to pixel noise. In last level, CCL algorithm is used for sorts out foregrounds pixels are grouped into meaningful connected regions and their features.

scholarly journals Background Modelling using a Q-Tree Based Foreground Segmentation

2020 ◽  
Vol 21 (1) ◽  
pp. 17-31
Author(s):  
S Shahidha Banu ◽  
N Maheswari
Keyword(s):  
Background Subtraction ◽  
Moving Objects ◽  
False Negative ◽  
Region Of Interest ◽  
Intensity Variation ◽  
Experimental Results ◽  
Foreground Segmentation ◽  
Foreground Detection ◽  
Foreground Object ◽  
Background Modelling

Background modelling is an empirical part in the procedure of foreground mining of idle and moving objects. The foreground object detection has become a challenging phenomenon due to intermittent objects, intensity variation, image artefact and dynamic background in the video analysis and video surveillance applications. In the video surveillances application, a large amount of data is getting processed by everyday basis. Thus it needs an efficient background modelling technique which could process those larger sets of data which promotes effective foreground detection. In this paper, we presented a renewed background modelling method for foreground segmentation. The main objective of the work is to perform the foreground extraction only inthe intended region of interest using proposed Q-Tree algorithm. At most all the present techniques consider their updates to the pixels of the entire frame which may result in inefficient foreground detection with a quick update to slow moving objects. The proposed method contract these defect by extracting the foreground object by controlling the region of interest (the region only where the background subtraction is to be performed) and thereby reducing the false positive and false negative. The extensive experimental results and the evaluation parameters of the proposed approach with the state of art method were compared against the most recent background subtraction approaches. Moreover, we use challenge change detection dataset and the efficiency of our method is analyzed in different environmental conditions (indoor, outdoor) from the CDnet2014 dataset and additional real time videos. The experimental results were satisfactorily verified the strengths and weakness of proposed method against the existing state-of-the-art background modelling methods.

Sign in / Sign up

Export Citation Format

Share Document