Weakly Supervised Foreground Object Detection Network Using Background Model Image
In visual surveillance, deep learning-based foreground object detection algorithms are superior to classical background subtraction (BGS)-based algorithms. However, deep learning-based methods are limited because detection performance deteriorates in a new environment different from the training env...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9910159/ |
| _version_ | 1811239601421942784 |
|---|---|
| author | Jae-Yeul Kim Jong-Eun Ha |
| author_facet | Jae-Yeul Kim Jong-Eun Ha |
| author_sort | Jae-Yeul Kim |
| collection | DOAJ |
| description | In visual surveillance, deep learning-based foreground object detection algorithms are superior to classical background subtraction (BGS)-based algorithms. However, deep learning-based methods are limited because detection performance deteriorates in a new environment different from the training environment. This limitation can be solved by retraining the model using additional ground-truth labels in the new environment. However, generating ground-truth labels for visual surveillance is time-consuming and expensive. This paper proposes a method that does not require foreground labels when adapting to a new environment. To this end, we propose an integrated network that produces two kinds of outputs a background model image and a foreground object map. We can adapt to the new environment by retraining using a background model image. The proposed method consists of one encoder and two decoders for detecting foreground objects and a background model image. It is designed to enable real-time processing with desktop GPUs. The proposed method shows 14.46% improved FM in a new environment different from training and 11.49% higher FM than the latest BGS algorithm. |
| first_indexed | 2024-04-12T13:02:50Z |
| format | Article |
| id | doaj.art-b71acd45be0f4b54bdeb61ad0d0bb648 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-12T13:02:50Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-b71acd45be0f4b54bdeb61ad0d0bb6482022-12-22T03:32:07ZengIEEEIEEE Access2169-35362022-01-011010572610573310.1109/ACCESS.2022.32119879910159Weakly Supervised Foreground Object Detection Network Using Background Model ImageJae-Yeul Kim0https://orcid.org/0000-0002-7765-4972Jong-Eun Ha1https://orcid.org/0000-0002-4144-1000Graduate School of Information and Communication Engineering, Daegu Gyeongbuk Institute of Science and Technology (DGIST), Daegu, South KoreaDepartment of Mechanical and Automotive Engineering, Seoul National University of Science and Technology, Seoul, South KoreaIn visual surveillance, deep learning-based foreground object detection algorithms are superior to classical background subtraction (BGS)-based algorithms. However, deep learning-based methods are limited because detection performance deteriorates in a new environment different from the training environment. This limitation can be solved by retraining the model using additional ground-truth labels in the new environment. However, generating ground-truth labels for visual surveillance is time-consuming and expensive. This paper proposes a method that does not require foreground labels when adapting to a new environment. To this end, we propose an integrated network that produces two kinds of outputs a background model image and a foreground object map. We can adapt to the new environment by retraining using a background model image. The proposed method consists of one encoder and two decoders for detecting foreground objects and a background model image. It is designed to enable real-time processing with desktop GPUs. The proposed method shows 14.46% improved FM in a new environment different from training and 11.49% higher FM than the latest BGS algorithm.https://ieeexplore.ieee.org/document/9910159/Visual surveillanceweakly superviseddeep learningforeground object detection |
| spellingShingle | Jae-Yeul Kim Jong-Eun Ha Weakly Supervised Foreground Object Detection Network Using Background Model Image IEEE Access Visual surveillance weakly supervised deep learning foreground object detection |
| title | Weakly Supervised Foreground Object Detection Network Using Background Model Image |
| title_full | Weakly Supervised Foreground Object Detection Network Using Background Model Image |
| title_fullStr | Weakly Supervised Foreground Object Detection Network Using Background Model Image |
| title_full_unstemmed | Weakly Supervised Foreground Object Detection Network Using Background Model Image |
| title_short | Weakly Supervised Foreground Object Detection Network Using Background Model Image |
| title_sort | weakly supervised foreground object detection network using background model image |
| topic | Visual surveillance weakly supervised deep learning foreground object detection |
| url | https://ieeexplore.ieee.org/document/9910159/ |
| work_keys_str_mv | AT jaeyeulkim weaklysupervisedforegroundobjectdetectionnetworkusingbackgroundmodelimage AT jongeunha weaklysupervisedforegroundobjectdetectionnetworkusingbackgroundmodelimage |