In this paper, we propose a privacy-preserving semantic segmentation method that uses encrypted images and models with the vision transformer (ViT), called the segmentation transformer (SETR). The combined use of encrypted images and SETR allows us not only to apply images without sensitive visual information to SETR as query images but to also maintain the same accuracy as that of using plain images. Previously, privacy-preserving methods with encrypted images for deep neural networks have focused on image classification tasks. In addition, the conventional methods result in a lower accuracy than models trained with plain images due to the influence of image encryption. To overcome these issues, a novel method for privacy-preserving semantic segmentation is proposed by using an embedding that the ViT structure has for the first time. In experiments, the proposed privacy-preserving semantic segmentation was demonstrated to have the same accuracy as that of using plain images under the use of encrypted images.

End-to-end systems have demonstrated state-of-the-art performance on many tasks related to automatic speech recognition (ASR) and dialect identification (DID). In this paper, we propose multi-task learning of Japanese DID and multi-dialect ASR (MD-ASR) systems with end-to-end models. Since Japanese dialects have variety in both linguistic and acoustic aspects of each dialect, Japanese DID requires simultaneously considering linguistic and acoustic features. One solution realizing Japanese DID using these features is to use transcriptions from ASR when performing DID. However, transcribing Japanese multi-dialect speech into text is regarded as a challenging task in ASR because there are big gaps in linguistic and acoustic features between a dialect and standard Japanese. One solution is dialect-aware ASR modeling, which means DID is performed with ASR. Therefore, the multi-task learning framework of Japanese DID and ASR is proposed to represent the dependency of them. We explore three systems as part of the proposed framework, changing the order in which DID and ASR are performed. In the experiments, Japanese multi-dialect ASR and DID tests were conducted on our home-made Japanese multi-dialect database and a standard Japanese database. The proposed transformer-based systems

Japan's largest platform for academic e-journals: J-STAGE is a full text database for reviewed academic papers published by Japanese societies

SNS providers are known to carry out the recompression and resizing of uploaded images, but most conventional methods for detecting fake images/tampered images are not robust enough against such operations. In this paper, we propose a novel method for detecting fake images, including distortion caused by image operations such as image compression and resizing. We select a robust hashing method, which retrieves images similar to a query image, for fake-image/tampered-image detection, and hash values extracted from both reference and query images are used to robustly detect fake-images for the first time. If there is an original hash code from a reference image for comparison, the proposed method can more robustly detect fake images than conventional methods. One of the practical applications of this method is to monitor images, including synthetic ones sold by a company. In experiments, the proposed fake-image detection is demonstrated to outperform state-of-the-art methods under the use of various datasets including fake images generated with GANs.

Deep neural networks (DNNs) are well known to be vulnerable to adversarial examples (AEs). In previous studies, the use of models encrypted with a secret key was demonstrated to be robust against white-box attacks, but not against black-box ones. In this paper, we propose a novel method using the vision transformer (ViT) that is a random ensemble of encrypted models for enhancing robustness against both white-box and black-box attacks. In addition, a benchmark attack method, called AutoAttack, is applied to models to test adversarial robustness objectively. In experiments, the method was demonstrated to be robust against not only white-box attacks but also black-box ones in an image classification task on the CIFAR-10 and ImageNet datasets. The method was also compared with the state-of-the-art in a standardized benchmark for adversarial robustness, RobustBench, and it was verified to outperform conventional defenses in terms of clean accuracy and robust accuracy.

This paper proposes a novel pseudo multi-exposure image fusion based on a single image. Multi-exposure image fusion is a method to produce images without saturation regions, by using photos with different exposures. However, it is difficult to take photos suited for the multi-exposure image fusion when we take a photo of dynamic scenes or record a video. In addition, the multi-exposure image fusion cannot be applied to existing images with a single exposure or videos. The proposed method enables us to produce pseudo multi-exposure images from a single image. To produce multi-exposure images, the proposed method utilizes a relationship between exposure values and pixel values. The relationship is obtained by assuming that a digital camera has a linear response function. Moreover, it is shown that the use of a local contrast enhancement method allows us to produce pseudo multi-exposure images with higher quality. Most of conventional multi-exposure image fusion methods are also applicable to the pseudo multi-exposure images. Experimental results show the effectiveness of the proposed method by comparing the proposed one with conventional ones.