Fake Colorized Image Detection with Channel-wise Convolution based Deep-learning Framework

Long, Zhuo; Tan, Shunquan; Zeng, Jishen; Lit, Bin

doi:10.23919/apsipa.2018.8659761

Cited by 16 publications

(13 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Size of the images are (3456 × 4608) or (4608 × 3456) pixels. For every original image, 10 to 11 tampered images (i.e., with copy-move [3,4], cut-paste [2], retouching [5,6] and colorizing [7,8]) are obtained.…”

Section: Experimental Design Materials and Methodsmentioning

confidence: 99%

A dataset of 1050-tampered color and grayscale images (CG-1050)

2020

View full text Add to dashboard Cite

This paper presents the CG-1050 dataset consisting of 100 original images, 1050 tampered images and their corresponding masks. The dataset is organized into four directories: original images, tampered images, mask images, and a description file. The directory of original images includes 15 color and 85 grayscale images. The directory of tampered images has 1050 images obtained through one of the following type of tampering: copy-move, cut-paste, retouching, and colorizing. The true mask between every pair of original and its tampered image is included in the mask directory (1380 masks). The description file shows the names of the images (i.e., original, tampered and mask), the image description, the photo location, the type of tampering, and the manipulated object in the image. With this dataset, the researchers can train and validate fake image classification methods, either for labelling the tampered image or for forgery pixel-detection.

show abstract

Section: Experimental Design Materials and Methodsmentioning

confidence: 99%

A dataset of 1050-tampered color and grayscale images (CG-1050)

2020

View full text Add to dashboard Cite

show abstract

“…For this specific forensic problem of colorized image detection, we have some observations about data and network: CIs shared by the authors of [6] and used in [6,15] are in a lossless format without compression on the artificially generated color information, and NIs from ImageNet are in the lossy JPEG format; in the meanwhile, the weights of first layer of WISERNet are initialized with SRM residual filters [5] and untrainable 2 . Therefore, it is natural to raise the following question: Does WISERNet, as used in [15], rather capture the difference of processing history between NIs and CIs (i.e., JPEG compressed or not), or the desired "essential" color difference?…”

Section: Data Preparation and Network 21 Motivationmentioning

confidence: 99%

“…We construct two sets of data and the only difference is whether CIs are JPEG compressed or not. Following [6] and [15], three stateof-the-art colorization algorithms (Ma [8], Mb [13], and Mc [7]) are adopted for producing CIs. NIs come from ImageNet dataset [3].…”

Section: Data Preparationmentioning

confidence: 99%

“…It is indeed difficult to distinguish, with naked human eyes, which images are artificially colorized. Accordingly, distinguishing between natural images (NIs) and colorized images (CIs) has drawn increasing interest among the image forensics research community [6,15].…”

Section: Introductionmentioning

confidence: 99%

“…Then, they trained the support vector machine (SVM) classifiers to identify fake colorized images. Zhuo et al [15] achieved the state-of-the-art detection performance on the experimental database shared by [6], by making use of an advanced CNN-based color image steganalyzer called WISERNet (WIder SEparate-then-Reunion Network) [12]. WISERNet is particularly good at detecting weak signals in images and its first layer has thirty 5 × 5 residual filters from the well-known steganalytic Spatial Rich Model (SRM) [5].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Impact of Data Preparation and CNN’s First Layer on Performance of Image Forensics: A Case Study of Detecting Colorized Images

Quan

Wang

Yan

et al. 2019

IEEE/WIC/ACM International Conference on Web Intelligence - Companion Volume

View full text Add to dashboard Cite

In the field of image forensics, many convolutional neural network (CNN)-based forensic methods have been proposed and generally achieved the state-of-the-art performance. However, some questions are worth studying and answering regarding the trustworthiness of such methods, including for example the appropriateness of the discriminative information automatically extracted by CNN and the generalization performance on "unseen" data during the testing phase. In this paper, we study these questions in the case of a specific forensic problem of distinguishing between natural images (NIs) and colorized images (CIs). Through a series of experiments, we analyze the impact of data preparation and setting of the first layer of a recent state-of-the-art CNN-based method on the detector's forensic performance, in particular the generalization capability. We obtain some interesting observations which can serve as useful hints for carrying out image forensics experiments. Moreover, we propose a very simple method to improve the generalization performance of colorized image detection by combining decision results from CNN models with different settings at the network's first layer.

show abstract

Channel Difference Based Regeneration Architecture for Fake Colorized Image Detection

Phutke

Murala²

2022

Communications in Computer and Information Science

View full text Add to dashboard Cite

Fake Colorized Image Detection with Channel-wise Convolution based Deep-learning Framework

Cited by 16 publications

References 6 publications

A dataset of 1050-tampered color and grayscale images (CG-1050)

A dataset of 1050-tampered color and grayscale images (CG-1050)

Impact of Data Preparation and CNN’s First Layer on Performance of Image Forensics: A Case Study of Detecting Colorized Images

Channel Difference Based Regeneration Architecture for Fake Colorized Image Detection

Contact Info

Product

Resources

About