A Deep Learning Approach in the DCT Domain to Detect the Source of HDR Images
Although high dynamic range (HDR) is now a common format of digital images, limited work has been done for HDR source forensics. This paper presents a method based on a convolutional neural network (CNN) to detect the source of HDR images, which is built in the discrete cosine transform (DCT) domain...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-12-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/9/12/2053 |
| _version_ | 1797545882413957120 |
|---|---|
| author | Jiayu Wang Hongquan Wang Xinshan Zhu Pengwei Zhou |
| author_facet | Jiayu Wang Hongquan Wang Xinshan Zhu Pengwei Zhou |
| author_sort | Jiayu Wang |
| collection | DOAJ |
| description | Although high dynamic range (HDR) is now a common format of digital images, limited work has been done for HDR source forensics. This paper presents a method based on a convolutional neural network (CNN) to detect the source of HDR images, which is built in the discrete cosine transform (DCT) domain. Specifically, the input spatial image is converted into DCT domain with discrete cosine transform. Then, an adaptive multi-scale convolutional (AMSC) layer extracts features related to HDR source forensics from different scales. The features extracted by AMSC are further processed by two convolutional layers with pooling and batch normalization operations. Finally, classification is conducted by a fully connected layer with Softmax function. Experimental results indicate that the proposed DCT-CNN outperforms the state-of-the-art schemes, especially in accuracy, robustness, and adaptability. |
| first_indexed | 2024-03-10T14:22:14Z |
| format | Article |
| id | doaj.art-3ff96cb602c840758f4e81a80fc148ea |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-10T14:22:14Z |
| publishDate | 2020-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-3ff96cb602c840758f4e81a80fc148ea2023-11-20T23:18:33ZengMDPI AGElectronics2079-92922020-12-01912205310.3390/electronics9122053A Deep Learning Approach in the DCT Domain to Detect the Source of HDR ImagesJiayu Wang0Hongquan Wang1Xinshan Zhu2Pengwei Zhou3School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, ChinaSchool of Electrical and Information Engineering, Tianjin University, Tianjin 300072, ChinaSchool of Electrical and Information Engineering, Tianjin University, Tianjin 300072, ChinaNational Laboratory of Industrial Control Technology, College of Control Science and Engineering, Zhejiang University, Hangzhou 310027, ChinaAlthough high dynamic range (HDR) is now a common format of digital images, limited work has been done for HDR source forensics. This paper presents a method based on a convolutional neural network (CNN) to detect the source of HDR images, which is built in the discrete cosine transform (DCT) domain. Specifically, the input spatial image is converted into DCT domain with discrete cosine transform. Then, an adaptive multi-scale convolutional (AMSC) layer extracts features related to HDR source forensics from different scales. The features extracted by AMSC are further processed by two convolutional layers with pooling and batch normalization operations. Finally, classification is conducted by a fully connected layer with Softmax function. Experimental results indicate that the proposed DCT-CNN outperforms the state-of-the-art schemes, especially in accuracy, robustness, and adaptability.https://www.mdpi.com/2079-9292/9/12/2053image forensicshigh dynamic rangeinverse tone mappingdiscrete cosine transformconvolutional neural networks |
| spellingShingle | Jiayu Wang Hongquan Wang Xinshan Zhu Pengwei Zhou A Deep Learning Approach in the DCT Domain to Detect the Source of HDR Images Electronics image forensics high dynamic range inverse tone mapping discrete cosine transform convolutional neural networks |
| title | A Deep Learning Approach in the DCT Domain to Detect the Source of HDR Images |
| title_full | A Deep Learning Approach in the DCT Domain to Detect the Source of HDR Images |
| title_fullStr | A Deep Learning Approach in the DCT Domain to Detect the Source of HDR Images |
| title_full_unstemmed | A Deep Learning Approach in the DCT Domain to Detect the Source of HDR Images |
| title_short | A Deep Learning Approach in the DCT Domain to Detect the Source of HDR Images |
| title_sort | deep learning approach in the dct domain to detect the source of hdr images |
| topic | image forensics high dynamic range inverse tone mapping discrete cosine transform convolutional neural networks |
| url | https://www.mdpi.com/2079-9292/9/12/2053 |
| work_keys_str_mv | AT jiayuwang adeeplearningapproachinthedctdomaintodetectthesourceofhdrimages AT hongquanwang adeeplearningapproachinthedctdomaintodetectthesourceofhdrimages AT xinshanzhu adeeplearningapproachinthedctdomaintodetectthesourceofhdrimages AT pengweizhou adeeplearningapproachinthedctdomaintodetectthesourceofhdrimages AT jiayuwang deeplearningapproachinthedctdomaintodetectthesourceofhdrimages AT hongquanwang deeplearningapproachinthedctdomaintodetectthesourceofhdrimages AT xinshanzhu deeplearningapproachinthedctdomaintodetectthesourceofhdrimages AT pengweizhou deeplearningapproachinthedctdomaintodetectthesourceofhdrimages |