The growing diversity of digital face manipulation techniques has led to an urgent need for a universal and robust detection technology to mitigate the risks posed by malicious forgeries. We present a blended-based detection approach that has robust applicability to unseen datasets. It combines a method for generating synthetic training samples, i.e., reconstructed blended images, that incorporate potential deepfake generator artifacts and a detection model, a multi-scale feature reconstruction network, for capturing the generic boundary artifacts and noise distribution anomalies brought about by digital face manipulations. Experiments demonstrated that this approach results in better performance in both cross-manipulation detection and cross-dataset detection on unseen data.
With the growing use of camera devices, the industry has many image datasets that provide more opportunities for collaboration between the machine learning community and industry. However, the sensitive information in the datasets discourages data owners from releasing these datasets. Despite recent research devoted to removing sensitive information from images, they provide neither meaningful privacy-utility trade-off nor provable privacy guarantees. In this study, with the consideration of the perceptual similarity, we propose perceptual indistinguishability (PI) as a formal privacy notion particularly for images. We also propose PI-Net, a privacy-preserving mechanism that achieves image obfuscation with PI guarantee. Our study shows that PI-Net achieves significantly better privacy utility trade-off through public image data.
Randomized smoothing has established state-of-the-art provable robustness against $\ell_2$ norm adversarial attacks with high probability. However, the introduced Gaussian data augmentation causes a severe decrease in natural accuracy. We come up with a question, "Is it possible to construct a smoothed classifier without randomization while maintaining natural accuracy?". We find the answer is definitely yes. We study how to transform any classifier into a certified robust classifier based on a popular and elegant mathematical tool, Bernstein polynomial. Our method provides a deterministic algorithm for decision boundary smoothing. We also introduce a distinctive approach of norm-independent certified robustness via numerical solutions of nonlinear systems of equations. Theoretical analyses and experimental results indicate that our method is promising for classifier smoothing and robustness certification.
In this paper, we reformulate the non-convex $\ell_q$-norm minimization problem with $q\in(0,1)$ into a 2-step problem, which consists of one convex and one non-convex subproblems, and propose a novel iterative algorithm called QISTA ($\ell_q$-ISTA) to solve the $\left(\ell_q\right)$-problem. By taking advantage of deep learning in accelerating optimization algorithms, together with the speedup strategy that using the momentum from all previous layers in the network, we propose a learning-based method, called QISTA-Net-s, to solve the sparse signal reconstruction problem. Extensive experimental comparisons demonstrate that the QISTA-Net-s yield better reconstruction qualities than state-of-the-art $\ell_1$-norm optimization (plus learning) algorithms even if the original sparse signal is noisy. On the other hand, based on the network architecture associated with QISTA, with considering the use of convolution layers, we proposed the QISTA-Net-n for solving the image CS problem, and the performance of the reconstruction still outperforms most of the state-of-the-art natural images reconstruction methods. QISTA-Net-n is designed in unfolding QISTA and adding the convolutional operator as the dictionary. This makes QISTA-Net-s interpretable. We provide complete experimental results that QISTA-Net-s and QISTA-Net-n contribute the better reconstruction performance than the competing.
Binary embedding of high-dimensional data aims to produce low-dimensional binary codes while preserving discriminative power. State-of-the-art methods often suffer from high computation and storage costs. We present a simple and fast embedding scheme by first downsampling N-dimensional data into M-dimensional data and then multiplying the data with an MxM circulant matrix. Our method requires O(N +M log M) computation and O(N) storage costs. We prove if data have sparsity, our scheme can achieve similarity-preserving well. Experiments further demonstrate that though our method is cost-effective and fast, it still achieves comparable performance in image applications.
Template matching is widely used for many applications in image and signal processing and usually is time-critical. Traditional methods usually focus on how to reduce the search locations by coarse-to-fine strategy or full search combined with pruning strategy. However, the computation cost of those methods is easily dominated by the size of signal N instead of that of template K. This paper proposes a probabilistic and fast matching scheme, which computation costs requires O(N) additions and O(K \log K) multiplications, based on cross-correlation. The nuclear idea is to first downsample signal, which size becomes O(K), and then subsequent operations only involves downsampled signals. The probability of successful match depends on cross-correlation between signal and the template. We show the sufficient condition for successful match and prove that the probability is high for binary signals with K^2/log K >= O(N). The experiments shows this proposed scheme is fast and efficient and supports the theoretical results.