26 February 2019 in IEEE Transactions on Information Forensics and Security
We address the problem of data-driven image manipulation detection in the presence of an attacker with limited knowledge about the detector. Specifically, we assume that the attacker knows the architecture of the detector, the training data and the class of features V the detector can rely on. In order to get an advantage in his race of arms with the attacker, the analyst designs the detector by relying on a subset of features chosen at random in V. Given its ignorance about the exact feature set, the adversary attacks a version of the detector based on the entire feature set. In this way, the effectiveness of the attack diminishes since there is no guarantee that attacking a detector working in the full feature space will result in a successful attack against the reduced-feature detector. We theoretically prove that, thanks to random feature selection, the security of the detector increases significantly at the expense of a negligible loss of performance in the absence of attacks. We also provide an experimental validation of the proposed procedure by focusing on the detection of two specific kinds of image manipulations, namely adaptive histogram equalization and median filtering. The experiments confirm the gain in security at the expense of a negligible loss of performance in the absence of attacks.
7 July 2021 in IEEE Access
As a basic task in the field of computer vision, target detection has been concerned by many researchers. The performance of target detection method is also directly related to the research in many advanced semantic fields. Current general target detection methods are not effective in small target detection, so this paper studies the problem of small target detection and proposes a small target detection method based on deep learning with considerate feature and effectively expanded sample size. Firstly, according to the characteristics of convolutional neural network, we improve the current deep network architecture which performs well in target detection, and introduce considerate multi-feature and multi-scale detection. Then, we transform the high-resolution images obtained on the Internet by combining two groups of sampling method, so that the feature distribution of the high-resolution target is closer to that of the low-resolution target, thus effectively expanding the training data set, solving the problem that small target data is difficult to be labeled and effectively avoiding overfitting. The results show the effectiveness of the improved method in small target detection. In addition, in view of the shortage of small target detection review literature, this paper gives a comprehensive and detailed introduction to the field of small target detection in terms of related work and future work.
24 June 2022 in IEEE Access
A large number of forensics research focus on operation detection to reveal the evidence of forgery action in the digital image. In the early works, analyst firstly model the probability distribution of single operation, and design the forensic tools based on feature extraction and machine learning based classifier. With increasing dimension of the feature and facing multiple operations detection scenario, the physical meaning of the feature gradually become ambiguous. Especially, since deep learning algorithm was used in forensic research, the automatic feature selection and making decision with high performance of classification conceals the intrinsic forensic clues. In this paper, we explore the availability of feature for operation detection in the operation chain, so called forensicability. An anti-forensic attack algorithm is introduced to formulate the impact on the feature due to the following operation. We propose two measurements: attack angle and scale, mutual information scale to indicate the forensic feature variation after the image manipulated by the following operation. The uncoupled relationship can be revealed by our methods. In the experiments, four operation chains involving ten operations are considered as the case study. The results are encouraging and improve the explanation of the forensics method based on high dimensional features.
18 March 2025 in IEEE Internet of Things Journal
As unmanned robotic operations become prevalent in various fields, the presence of moving objects in complex scenes poses challenges to key technologies, such as environment mapping, obstacle avoidance, and trajectory prediction. In this article, we propose FP-MOS, a LiDAR-based dynamic object segmentation framework that integrates prediction information to continuously capture dynamic object information under constrained data conditions. First, a prediction module (PM) is employed to obtain predicted range images, enhancing the continuity of temporal information in the point cloud within the segmentation network. This improved continuity helps mitigate misjudgment issues caused by dynamic object turns and retrieves previously missed moving objects. Additionally, a motion attention weight guidance module is introduced, which accurately captures dynamic point cloud features through the transmission of point cloud weights between adjacent frames. Finally, we apply an improved adaptive local outlier factor (ALOF) method to filter out outliers in the segmentation results. Experimental results on the SemanticKITTI-MOS and Apollo datasets show that our method achieves leading Intersection-over-Union (IoU) scores of 79.3% and 79.0%, demonstrating the effectiveness of the FP-MOS method in network design and data optimization.