scholarly journals Dynamic Codebook for Foreground Segmentation in a Video

2017 ◽  
Vol 10 (2) ◽  
pp. 144-155
Author(s):  
Worapan Kusakunniran ◽  
Rawitas Krungkaew
Keyword(s):  
Computer Vision ◽  
Real World ◽  
Early Stage ◽  
Image Sequences ◽  
Foreground Segmentation ◽  
Background Information ◽  
Detection Algorithms ◽  
Change Detection Algorithms ◽  
Novel Method

The foreground segmentation in a video is a way to extract changes in image sequences. It is a key task in an early stage of many applications in the computer vision area. The information of changes in the scene must be segmented before any further analysis could be taken place. However, it remains with difficulties caused by several real-world challenges such as cluttered backgrounds, changes of the illumination, shadows, and long-term scene changes. This paper proposes a novel method, namely a dynamic codebook (DCB), to address such challenges of the dynamic backgrounds. It relies on a dynamic modeling of the background scene. Initially, a codebook is constructed to represent the background information of each pixel over a period of time. Then, a dynamic boundary of the codebook will be made to support variations of the background. The revised codebook will always be adaptive to the new background's environments. This makes the foreground segmentation more robust to the changes of background scene. The proposed method has been evaluated by using the changedetection.net (CDnet) benchmark which is a well-known video dataset for testing change-detection algorithms. The experimental results and comprehensive comparisons have shown a very promising performance of the proposed method.

scholarly journals Robust Visual Tracking Based on Adaptive Convolutional Features and Offline Siamese Tracker

Sensors ◽  
2018 ◽  
Vol 18 (7) ◽  
pp. 2359 ◽  
Author(s):  
Ximing Zhang ◽  
Mingang Wang
Keyword(s):  
Visual Tracking ◽  
Training Data ◽  
Background Information ◽  
Appearance Model ◽  
Dimensionality Reduction Technique ◽  
Wide Range ◽  
Model Update ◽  
Tracking Ability ◽  
Novel Method

Robust and accurate visual tracking is one of the most challenging computer vision problems. Due to the inherent lack of training data, a robust approach for constructing a target appearance model is crucial. The existing spatially regularized discriminative correlation filter (SRDCF) method learns partial-target information or background information when experiencing rotation, out of view, and heavy occlusion. In order to reduce the computational complexity by creating a novel method to enhance tracking ability, we first introduce an adaptive dimensionality reduction technique to extract the features from the image, based on pre-trained VGG-Net. We then propose an adaptive model update to assign weights during an update procedure depending on the peak-to-sidelobe ratio. Finally, we combine the online SRDCF-based tracker with the offline Siamese tracker to accomplish long term tracking. Experimental results demonstrate that the proposed tracker has satisfactory performance in a wide range of challenging tracking scenarios.

Tower Design, Build and Test as a STEAM Project

2020 ◽  
pp. 154-169
Author(s):  
Judith Bazler
Keyword(s):  
Real World ◽  
Background Information ◽  
Engineering Practice ◽  
School Students ◽  
Final Project ◽  
Graph Paper ◽  
Design Build ◽  
Cooperative Team ◽  
And Mathematics

The next generation science standards promote the teaching of engineering skills including the designing, testing, and building of models. Tower building can yield real world experience that not only provides the student with physics and mathematics through motion and stability but also through the explanation of the use of models and the engineering practice of design, redesign, and testing of these models. Tia Pliskow used the project of building a tower with her middle school students in order to provide a cooperative team long-term project. She focused first on the design, using background information on existing towers. She required each team to design their tower first using graph paper and scale. This process stressed the need for science, technology, engineering, art, and mathematics. The case included in this article expands her process by including a cost analysis attempting to promote real world engineering, links to more content, and final project photos. In addition, by building a shake platform, a test for the tower is added.

Abstract 17109: Real World Assessment of Atrial Fibrillation and Other Arrhythmias After Cryptogenic Stroke

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_3) ◽  
Author(s):  
Naga Venkata K Pothineni ◽  
Uyanga Batnyam ◽  
Jeffrey Arkles ◽  
John Bullinga ◽  
Brett L CUCCHIARA ◽  
...  
Keyword(s):  
Atrial Fibrillation ◽  
Clinical Trials ◽  
Real World ◽  
Cryptogenic Stroke ◽  
Premature Ventricular Contractions ◽  
Entire Cohort ◽  
Detection Algorithms ◽  
Term Monitoring

Introduction: Long-term monitoring for atrial fibrillation (AF) is recommended in patients, who have experienced a cryptogenic stroke (CS). Clinical trials have identified AF in ~30% of patients after 3 years of continuous monitoring with insertable cardiac monitors (ICMs). Hypothesis: In a real-world analysis from a large academic healthcare system, we sought to evaluate a CS population with ICMs and a) determine the yield of AF and subsequent initiation of anticoagulation; and b) identify the presence of other arrhythmias. Methods: We evaluated all CS patients who had received an ICM between October 2014 and April 2020. We manually reviewed all stored electrocardiograms that were automatically labeled as AF by the ICM and adjudicated them as either a) AF or b) other cardiac arrhythmia including premature atrial contractions (PAC), premature ventricular contractions (PVC), supraventricular tachycardia (SVT), or nonsustained ventricular tachycardia (NSVT). Results: A total of 84 CS patients with ICMs were included: 51% men, mean age 63 years, and mean CHA 2 DS 2 -VASc 4.1. Over a median follow-up duration of 15.7 months, there were 34 patients (40% of the cohort) who did not have any AF alerts. In the remaining 50 patients, there were 960 stored electrograms that were adjudicated. Only 154 recordings from 16 patients (19% of the entire cohort) were adjudicated as AF. Oral anticoagulation was initiated in all these patients with adjudicated AF. The remaining tracings, which had been automatically categorized by the ICM as AF alerts, represented 34 patients (40% of the cohort). These patients had other arrhythmias including frequent PACs or PVCs, SVT, or NSVT. Conclusions: Compared to clinical trials, our real-world assessment suggests that the yield of AF following CS is lower - approximately 20%. Our findings highlight the importance for reviewing device tracings given the high rates of false positive for AF. Further research to refine AF detection algorithms in ICMs is needed.

The application of black-box modeling in long-term tunnel deformation monitoring

2021 ◽  
Author(s):  
Kristof Maes ◽  
Stijn François ◽  
Wim Salens ◽  
Gerrit Feremans ◽  
Koen Segher
Keyword(s):  
Asset Management ◽  
Early Stage ◽  
Novelty Detection ◽  
Warning System ◽  
Principal Component ◽  
Black Box ◽  
Deformation Monitoring ◽  
Detection Algorithms ◽  
Tunnel Monitoring

<p>Tunnel closures related to maintenance and reconstruction works can lead to large economical costs and should therefore be avoided. This paper explores the use of novelty detection algorithms for long-term tunnel monitoring. The aim is to detect tunnel damage in an early stage, as such providing a tool to support the asset management. The proposed strategy is applied to the monitoring of the Waasland tunnel in Antwerp, where the deformations and temperatures have been monitored over a period of 14 months. The case demonstrates that novelty detection by means of principal component analysis enables the identification of minor changes in the tunnel response, and can therefore be embedded in an early detection warning system.</p>

scholarly journals Improving Deep Object Detection Algorithms for Game Scenes

Electronics ◽  
2021 ◽  
Vol 10 (20) ◽  
pp. 2527
Author(s):  
Minji Jung ◽  
Heekyung Yang ◽  
Kyungha Min
Keyword(s):  
Computer Vision ◽  
Object Detection ◽  
Real World ◽  
Computer Games ◽  
Recognition Performance ◽  
Role Playing ◽  
Scene Analysis ◽  
Research Topics ◽  
The Real ◽  
Detection Algorithms

The advancement and popularity of computer games make game scene analysis one of the most interesting research topics in the computer vision society. Among the various computer vision techniques, we employ object detection algorithms for the analysis, since they can both recognize and localize objects in a scene. However, applying the existing object detection algorithms for analyzing game scenes does not guarantee a desired performance, since the algorithms are trained using datasets collected from the real world. In order to achieve a desired performance for analyzing game scenes, we built a dataset by collecting game scenes and retrained the object detection algorithms pre-trained with the datasets from the real world. We selected five object detection algorithms, namely YOLOv3, Faster R-CNN, SSD, FPN and EfficientDet, and eight games from various game genres including first-person shooting, role-playing, sports, and driving. PascalVOC and MS COCO were employed for the pre-training of the object detection algorithms. We proved the improvement in the performance that comes from our strategy in two aspects: recognition and localization. The improvement in recognition performance was measured using mean average precision (mAP) and the improvement in localization using intersection over union (IoU).

MOTION ANALYSIS AND REPRESENTATION IN COMPUTER VISION

1993 ◽  
Vol 03 (04) ◽  
pp. 797-831 ◽  
Author(s):  
V. CAPPELLINI ◽  
A. MECOCCI ◽  
A. DEL BIMBO
Keyword(s):  
Computer Vision ◽  
Motion Analysis ◽  
Image Sequences ◽  
Motion Field ◽  
Short Term ◽  
High Interest ◽  
Motion Parameters ◽  
Qualitative Estimation

Motion analysis is of high interest in many different fields for a number of crucial applications. Short-term motion analysis addresses the computation of motion parameters or the qualitative estimation of the motion field. Long-term motion analysis aims at the understanding of motion and includes reasoning on motion properties. Image sequences are in general processed to perform the above motion analysis. These subjects are considered in this review with reference to the more significant results in the literature both at theory and application levels.

Tower Design as a STEAM Project

2017 ◽  
pp. 206-219
Author(s):  
Judith A. Bazler
Keyword(s):  
Real World ◽  
Background Information ◽  
Engineering Practice ◽  
Science Standards ◽  
School Students ◽  
Graph Paper ◽  
Cooperative Team ◽  
And Mathematics ◽  
Use Of Models

The next generation science standards promote the teaching of engineering skills including the designing, testing and building of models. Tower building can provide a real world experience that not only provides the students with physics and mathematics through motion and stability but the explanation of the use of models and the engineering practice of design, redesign and testing of these models. Tia Pilskow (2014) used the project of building a tower with her middle school students in order to provide a cooperative team long term project. She focused first on the design using background information on existing towers. She required each team to design their tower first using graph paper and scale. This process stressed the need for Art and Mathematics in the STEAM project. The science, technology and engineering also played a major part in the design. The case included in this article expands her process by including a cost analysis attempting to promote real world engineering.

Sign in / Sign up

Export Citation Format

Share Document