Median filtering is used as an anti-forensic technique to erase processing history of some image manipulations such as JPEG, resampling, etc. Thus, various detectors have been proposed to detect median filtered images. To counter these techniques, several anti-forensic methods have been devised as well. However, restoring the median filtered image is a typical ill-posed problem, and therefore it is still difficult to reconstruct the image visually close to the original image. Also, it is further hard to make the restored image have the statistical characteristic of the raw image for the anti-forensic purpose. To solve this problem, we present a median filtering anti-forensic method based on deep convolutional neural networks (CNNs), which can effectively remove traces from median filtered images. We adopt the framework of generative adversarial networks (GANs) to generate images that follow the underlying statistics of unaltered images, significantly enhancing forensic undetectability. Through extensive experiments, we demonstrate that our method successfully deceives the existing median filtering forensic techniques.

As technological developments have enabled high-quality fingerprint scanning, sweat pores, one of the Level 3 features of fingerprints, have been successfully used in automatic fingerprint recognition systems (AFRS). Since the pore extraction process is a critical step for AFRS, high accuracy is required. However, it is difficult to extract the pore correctly because the pore shape depends on the person, region, and pore type. To solve the problem, we have presented a pore extraction method using deep convolutional neural networks and pore intensity refinement. The deep networks are used to detect pores in detail using a large area of a fingerprint image. We then refine the pore information by finding local maxima to identify pores with different intensities in the fingerprint image. The experimental results show that our pore extraction method performs better than the state-of-the-art methods.

Image steganalysis is the task of detecting a secret message hidden in an image. Deep steganalysis using end-to-end deep learning has been successful in recent years, but previous studies focused on improving detection performance rather than designing a lightweight model for practical applications. This caused a deep steganalysis model to be heavy and computationally costly, making the model infeasible to deploy in real-world applications. To address this issue, we study an effective model design strategy for lightweight image steganalysis. Considering the domain-specific characteristics of steganalysis, we propose a simple yet effective block removal strategy that progressively removes a sequence of blocks from deep classification networks. This method involves the gradual removal of convolutional neural network blocks, starting from deeper ones. By doing so, the number of parameters and FLOPs are decreased without compromising the detection performance. Experimental results show that our removal strategy makes the EfficientNet-B0 variants 9.58 \(\times\) smaller and has 2.16 \(\times\) fewer FLOPs than the baseline while retaining detection accuracy of 90.73% and 82.40% that are on par with the baseline on BOSSBase and ALASKA#2 datasets, respectively. Backed by our in-depth analyses, the results indicate that only a few early layers are sufficient for effective image steganalysis.

Steganalysis refers to the study of identifying hidden messages in images inserted by steganography. Although detection performance is greatly improved when adopting convolutional neural networks (CNNs), they require sophisticated tricks, such as preprocessing for suppression of image content, using absolute and truncated activation functions, and utilizing domain knowledge. These tricks help networks train stably and mitigate the convergence problem of early stages in training, but they also restrict the flexibility of CNNs, which limits their performance. In this paper, we propose a local-source enhanced residual network (LSER) with end-to-end learning. The LSER is simple in its architecture but has two distinct characteristics from previous methods. First, the LSER uses residual blocks without any normalization. We find batch normalization is an unnecessary module in our framework. Second, a local-source skip connection is added to bypass features of different levels, which allows more abundant feature representation. Moreover, the LSER exhibits state-of-the-art results compared with the existing work in both spatial and JPEG domain steganalysis. Furthermore, a simple self-ensemble method further improves its performance without any side information.

The current research direction in generative models, such as the recently developed GPT4, aims to find relevant knowledge information for multimodal and multilingual inputs to provide answers. Under these research circumstances, the demand for multilingual evaluation of visual question answering (VQA) tasks, a representative task of multimodal systems, has increased. Accordingly, we propose a bilingual outside-knowledge VQA (BOK-VQA) dataset in this study that can be extended to multilingualism. The proposed data include 17K images, 17K question-answer pairs for both Korean and English and 280K instances of knowledge information related to question-answer content. We also present a framework that can effectively inject knowledge information into a VQA system by pretraining the knowledge information of BOK-VQA data in the form of graph embeddings. Finally, through in-depth analysis, we demonstrated the actual effect of the knowledge information contained in the constructed training data on VQA.

Face anti-spoofing (FAS) techniques are crucial for safeguarding face recognition systems against an ever-evolving landscape of spoofing attacks. While existing methods have made strides in detecting known attacks, robustness against unknown, potentially more sophisticated attack types remains a critical and unresolved challenge in the field. In this paper, we propose a novel framework integrating token-wise asymmetric contrastive learning with angular margin loss to enhance the robustness of FAS models against unknown attack types. The key idea to resolve the problem is to learn a feature space where live face features are densely distributed, whereas spoof face features are more dispersed. This is achieved through two novel strategies: (1) asymmetric contrastive learning, encouraging the FAS model to learn a compact distribution for live face features while relaxing the constraint on the distribution of spoof face features, and (2) token-wise learning, focusing on capturing intrinsic liveness cues from local region rather than identity or facial-related features. Additionally, an angular margin loss is incorporated to enhance the discriminative power of the learned features. Extensive experiments on public benchmark datasets demonstrate the superiority of our FAS model over state-of-the-art methods in cross-attack scenarios, showcasing its strong robustness to unknown attacks while maintaining unseen domain generalization capability.