Deep learning-based scene-awareness approach for intelligent change detection in videos

Chan, Yi-Tung

doi:10.1117/1.jei.28.1.013038

Cited by 10 publications

(3 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Through another approach, Zhang et al [147] used a Stacked Denoising Auto-Encoder (SDAE) to learn robust spatial features and modeled the background with density analysis, whereas Shafiee et al [145] employed Neural Reponse Mixture (NeREM) to learn deep features used in the Mixture of Gaussians (MOG) model [49]. In 2019, Chan [149] proposed a deep learning-based scene-awareness approach for change detection in video sequences thus applying the suitable background subtraction algorithm for the corresponding type of challenges.…”

Section: Machine Learning Modelsmentioning

confidence: 99%

Moving Objects Detection with a Moving Camera: A Comprehensive Review

Chapel¹,

Bouwmans²

2020

Preprint

View full text Add to dashboard Cite

During about 30 years, a lot of research teams have worked on the big challenge of detection of moving objects in various challenging environments. First applications concern static cameras but with the rise of the mobile sensors studies on moving cameras have emerged over time. In this survey, we propose to identify and categorize the different existing methods found in the literature. For this purpose, we propose to classify these methods according to the choose of the scene representation: one plane or several parts. Inside these two categories, the methods are grouped according to eight different approaches: panoramic background subtraction, dual cameras, motion compensation, subspace segmentation, motion segmentation, plane+parallax, multi planes and split image in blocks. A reminder of methods for static cameras is provided as well as the challenges with both static and moving cameras. Publicly available datasets and evaluation metrics are also surveyed in this paper.

show abstract

Section: Machine Learning Modelsmentioning

confidence: 99%

Moving Objects Detection with a Moving Camera: A Comprehensive Review

Chapel¹,

Bouwmans²

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Scene awareness has been studied using scene classification for choosing a change detection method suitable for the scene type [27]. However, they did not perform object detection.…”

Section: Introductionmentioning

confidence: 99%

Specialized indoor and outdoor scene-specific object detection models

Jamali,

Davidsson,

Khoshkangini

et al. 2024

Sixteenth International Conference on Machine Vision (ICMV 2023)

View full text Add to dashboard Cite

Object detection is a critical task in computer vision with applications across various domains, ranging from autonomous driving to surveillance systems. Despite extensive research on improving the performance of object detection systems, identifying all objects in different places remains a challenge. The traditional object detection approaches focus primarily on extracting and analyzing visual features without considering the contextual information about the places of objects. However, entities in many real-world scenarios closely relate to their surrounding environment, providing crucial contextual cues for accurate detection. This study investigates the importance and impact of places of images (indoor and outdoor) on object detection accuracy. To this purpose, we propose an approach that first categorizes images into two distinct categories: indoor and outdoor. We then train and evaluate three object detection models (indoor, outdoor, and general models) based on YOLOv5 and 19 classes of the PASCAL VOC dataset and 79 classes of COCO dataset that consider places. The experimental evaluations show that the specialized indoor and outdoor models have higher mAP (mean Average Precision) to detect objects in specific environments compared to the general model that detects objects found both indoors and outdoors. Indeed, the network can detect objects more accurately in similar places with common characteristics due to semantic relationships between objects and their surroundings, and the network's misdetection is diminished. All the results were analyzed statistically with t-tests.

show abstract

“…Digital image processing techniques began in the 1960s [8][9]. In 1964, an aerodynamic laboratory in California, USA, improved the images transmitted from space probes by means of digital image processing techniques, and the images of the Moon transmitted by the Prowler 7 satellite at that time were subjected to image processing by a computer, which succeeded in correcting the image distortion problems that had arisen in the space camera equipment [10][11][12]. Target recognition of images can be viewed as the process of distinguishing and extracting the same type of target from multiple types of targets, which can include not only recognizing and classifying between two types of targets but also dealing with the recognition between single independent targets [13][14].…”

Section: Introductionmentioning

confidence: 99%

Research on intelligent analysis and identification of visualization scenes in transport supervision hall based on image processing technology

Gu,

Hao,

Gong

et al. 2023

Applied Mathematics and Nonlinear Sciences

View full text Add to dashboard Cite

This paper firstly investigates the visual scene testing method with image processing technique and predicts the number of scenes by UML structure. Secondly, the scene recognition of the transport supervision hall is performed by using image processing technology, and the ant colony optimization algorithm is proposed for local search to update the scene information and edge extraction. Then, the ED-AlexNet network model is constructed to detect and identify the target scenes. Finally, an error matrix is introduced to calculate the confidence of the sample model distribution in the test set, and the recognition extraction performance and recognition accuracy of the ED-AlexNet network model are analyzed. The study shows that when the error matrix is introduced, the highest value of ED-AlexNet F – measure is close to 0.9, and the end value is over 160, which has a good performance of scene target recognition extraction. The average recognition accuracy of ED-AlexNet is higher than 95%, with good compatibility and high accuracy of recognition.

show abstract

Deep learning-based scene-awareness approach for intelligent change detection in videos

Cited by 10 publications

References 40 publications

Moving Objects Detection with a Moving Camera: A Comprehensive Review

Moving Objects Detection with a Moving Camera: A Comprehensive Review

Specialized indoor and outdoor scene-specific object detection models

Research on intelligent analysis and identification of visualization scenes in transport supervision hall based on image processing technology

Contact Info

Product

Resources

About