24 November 2020 in IEEE Access
Data analysis on edge/cloud computing systems is becoming more important. However, if the information being analyzed may leak personal identification, the owners tend to withhold or partially block the data to ensure privacy protection. The resulting data is often insufficiently detailed to permit useful analysis. As a result, the desired analysis accuracy may not be achieved. To deal with this issue, several studies have examined encryption systems that support data analysis of encrypted data. Among the likely encryption systems, we focus on those based on the random unitary transform. This is because the random unitary transform offers lower computational complexity than other encryption approaches, and several signal processing algorithms have been proposed. However, analysis models for distributed encrypted data, have not been studied deeply enough. In this paper, we construct an analysis model for random unitary transform encrypted data by deriving a sparse model based on the elastic-net solution. The analytical model can derive the same elastic-net solution as that yielded by processing the original data (i.e. without encryption). The analytical model supports distributed encryption, where different parts of a data set are encrypted at different sites independently. Moreover, the proposed encryption model solves the problem that the security strength of random unitary transform decreases under certain conditions. Data aggregation after distributed encryption improves the accuracy of analyzing distributed privacy-sensitive information.
8 August 2023 in Neural Networks
In image classification, a deep neural network (DNN) that is trained on undistorted images constitutes an effective decision boundary. Unfortunately, this boundary does not support distorted images, such as noisy or blurry ones, leading to accuracy drop-off. As a simple approach for classifying distorted images as well as undistorted ones, previous methods have optimized the trained DNN again on both kinds of images. However, in these methods, the decision boundary may become overly complicated during optimization because there is no regularization of the decision boundary. Consequently, this decision boundary limits efficient optimization. In this paper, we study a simple yet effective decision boundary for distorted image classification through the use of a novel loss, called a "neural activation pattern matching (NAPM) loss". The NAPM loss is based on recent findings that the decision boundary is a piecewise linear function, where each linear segment is constructed from a neural activation pattern in the DNN when an image is fed to it. The NAPM loss extracts the neural activation patterns when the distorted image and its undistorted version are fed to the DNN and then matches them with each other via the sigmoid cross-entropy. Therefore, it constrains the DNN to classify the distorted image and its undistorted version by the same linear segment. As a result, our loss accelerates efficient optimization by preventing the decision boundary from becoming overly complicated. Our experiments demonstrate that our loss increases the accuracy of the previous methods in all conditions evaluated.
9 November 2023 in IEEE Transactions on Multimedia
Deep learning has been widely studied for processing and understanding multimedia data, and it does help improve performance. Recent research has shown that deep models are vulnerable to images containing adverse weather corruptions, leading to a safety risk for numerous safety-critical systems ( e.g., autonomous driving systems). There are two problems with the current situation. First, collecting data under different weather scenarios is highly difficult in practice. Second, the performance degrades significantly when the training and test data are from different distributions, as exemplified by the weather corrupted test data. As a result, it is challenging to train a model without access to the images containing variations of various weather conditions, and it is difficult to make trained model generalized to unknown data under different weather conditions. In this paper, we introduce a Counterfactual Representation Learning (CRL) method to address these problems. Without access to training data including weather condition variations, our CRL makes the model resistant to unseen test data that has been corrupted by weather condition variations. Our basic idea is inspired by the perspective of counterfactual regularization. We build a causal model that introduces a counterfactual variable to eliminate the unobserved characteristics brought about by weather conditions. In particular, such a counterfactual variable is approximated by randomly shuffled features, echoing the previous empirical observation that the shuffling technique can perturb the shape details while preserving the local textures. We use information theoretic representation learning to encourage the neural networks to learn more powerful and robust features, which consist of two components. We conduct experiments on five benchmark datasets, namely, CIFAR-100-C, ImageNet-C, KITTI-C, BDD100k, and CityScapes-C, all of which contain weather corruption. The results of our experiments show that our proposed method can not only be a plug-and-play technique but also work nicely for both object recognition and detection.
17 April 2024 in IEEJ Transactions on Electrical and Electronic Engineering
Demand for use of indoor and outdoor surveillance cameras to prevent crime, analyze customer behavior in commercial facilities and other purposes has been increasing. For these types of applications, it is important to be able to estimate information that cannot be observed directly from the images, using the image data. One such type of information is the height of people appearing in the video. People's heights can be used, for example, to help identify a suspect in a crime investigation or as an attribute to understand customers better for marketing. However, many conventional technologies for estimating height from images have required special calibration, or the acquisition of other information in addition to the video. This paper proposes a method for estimating a person's height using only an image, without the need for any special calibration. The key points of the proposed method involve the following two processes. In the first process, a reference object of known size and the person are recognized and the person's height is estimated based on the relative sizes of the person and reference object in the image. The second process involves selecting frames from the video that are appropriate for use in the estimation process. In practical experiments using real video, the method was able to estimate people's heights with an average estimation error of 1.9%. © 2024 The Authors. IEEJ Transactions on Electrical and Electronic Engineering published by Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.
13 April 2026 in IEEE Transactions on Signal and Information Processing over Networks
We propose a denoising method for multimodal graph signals by an alternating minimization scheme that sequentially solves signal restoration and graph learning problems. Many complex-structured data, i.e., those on sensor networks, can capture multiple modalities at each measurement point, referred to as modalities. They are also assumed to have an underlying structure or correlations in the modality domain as well as in space. Such multimodal data are regarded as graph signals on a twofold graph and they are often corrupted by noise. Furthermore, their spatial/modality relationships are not always given a priori: We need to estimate twofold graphs during a denoising algorithm. In this paper, we consider a signal denoising method on twofold graphs, where graphs are learned simultaneously. Specifically, the graph learning subproblems are solved using the primal-dual splitting (PDS) algorithm, while the signal update has a closed-form solution. Parameters in this iterative algorithm are learned from training data by unrolling the iteration with deep algorithm unrolling. Experimental results on synthetic and real-world data demonstrate that the proposed method outperforms existing model- and deep learning-based graph signal denoising methods.
1 February 2024 in IEEJ Transactions on Electronics, Information and Systems
We consider design of sparse IIR filters based on the least absolute shrinkage and selection operator (LASSO) for reducing the computation cost of IIR filters. The design algorithm is summarized as follows. First, we solve a constrained quadratic programming problem where the constraint is the positive realness with maximum pole radius. Next, after fixing the denominator coefficients, we apply the LASSO to the filter design problem so that the numerator coefficients include zero. Finally, we solve a constrained quadratic programming with zero coefficients constraints. Moreover, we confirm the effectiveness of the proposed method by using two design examples.