Tetsushi Ohki

• Doctor of Engineering
• Professor (Associate) at Shizuoka University

In this study, we measure the Intrinsic Dimension (ID) of token embedding to estimate the intrinsic dimensions of the manifolds spanned by the representations, so as to evaluate their redundancy quantitatively compared to their extrinsic dimensionality. In detail, (1) we estimate the ID of token embeddings in small-scale language models and also mo...

This paper discusses the attack feasibility of Remote Adversarial Patch (RAP) targeting face detectors. The RAP that targets face detectors is similar to the RAP that targets general object detectors, but the former has multiple issues in the attack process the latter does not. (1) It is possible to detect objects of various scales. In particular,...

Multibiometrics, which uses multiple biometric traits to improve recognition performance instead of using only one biometric trait to authenticate individuals, has been investigated. Previous studies have combined individually acquired biometric traits or have not fully considered the convenience of the system.Focusing on a single face image, we pr...

Biometric recognition systems, known for their convenience, are widely adopted across various fields. However, their security faces risks depending on the authentication algorithm and deployment environment. Current risk assessment methods faces significant challenges in incorporating the crucial factor of attacker's motivation, leading to incomple...

Demographic bias is one of the major challenges for face recognition systems. The majority of existing studies on demographic biases are heavily dependent on specific demographic groups or demographic classifier, making it difficult to address performance for unrecognised groups. This paper introduces ``LabellessFace'', a novel framework that impro...

Social engineering involves manipulating a target to make it act to achieve the attacker's objectives. Currently, direct and short-term attacks are common in the mainstream. However, with the development of science and technology, there has been a shift toward extended and long-term attacks using VR and other technologies. We define advanced persis...

The number of insider threats and the expense of handling them increase every year, rendering it imperative to adopt measures against insider threats. In particular, insiders must adopt a psychological approach. Psychological approaches can be classified into three categories: deceiving insiders, demoralizing insiders, and luring insiders. The dece...

The war between cyber attackers and security analysts is gradually intensifying. Owing to the ease of obtaining and creating support tools, recent malware continues to diversify into variants and new species. This increases the burden on security analysts and hinders quick analysis. Identifying malware families is crucial for efficiently analyzing...

Privacy-preserving location synthesizers have been widely studied to perform private geo-data analysis. They have also been used for generating datasets for research or competitions. However, existing location synthesizers do not take into account the friendship information of users. Because friends tend to visit the same place at the same time in...

ID-less palmprint recognition is a biometric identification method using the pattern of the palm. It is highly available because it can obtain biometric information in a contactless manner using only a smartphone’s camera in everybody’s hands. It is highly convenient because it does not require an ID and uses only biometric information. Furthermore...

A secure machine learning technology that performs prediction while encrypting data using homomorphic encryption is being developed. However, Convolutional Neural Networks (CNN) on homomorphic encryption cannot use general non-linear activation functions, Thus, the classification accuracy is low. We proposed a novel method to create an activation f...

Biometric information can have high similarity among different people but high variability within the same person. Therefore, it is difficult to identify a person or distinguish between two people based on biometric information. One way to address these issues is to use certain biometric information as pivots to define the feature space of biometri...

CAPTCHA is a type of Turing test used to distinguish between humans and computing machine. However, image-based CAPTCHAs are losing their function as Turing tests owing to the improvement of image recognition using machine learning. This paper proposes an Adversarial CAPTCHA that provides attacking resistance to CAPTCHAs by using Adversarial Exampl...

Location traces are useful for various types of geo-data analysis tasks, and synthesizing location traces is a promising approach to geo-data analysis while protecting user privacy. However, existing location synthesizers do not consider friendship information of users. In particular, a co-location between friends is an important factor for synthes...

Nowadays, it is no longer uncommon for a PC to be infected with malware. User authentication based on only one legitimate credential (such as a password) may be insufficient for judging whether a user is legitimate. A typical solution to this problem is two-factor authentication, which is a method of authentication based on the presentation of two...

There is increasing recognition of the importance of social media in information transmission during disasters. However, it is difficult to distinguish between credible information and non-credible information from the deluge of content on social media. In order to ensure the transmission of highly-credible information during disasters, both the cr...

We have many opportunities to disclose privacy information in exchange for convenient services. In the context of privacy calculus, numerous research studies have been conducted to date on the relation between potential benefits and potential risks of privacy disclosure decisions [1, 2]. However, an unresolved problem in the privacy calculus is tha...

Micro nail authentication is an authentication mechanism that uses a minute part of the fingernail surface. Using a fingernail that grows over time, the right to be forgotten is satisfied because the biometric information is short-lived. The viability of micro nail authentication was proven with a prototype system in previous studies. However, ther...

Many smartphones are equipped with a biometric authentication function to prevent their unauthorized use. An authentication method using palmprint as a physical feature has been proposed. Palmprints can be authenticated without special devices because the features of palmprints can be acquired using a smartphone camera. However, impersonation may b...

In recent years, illegal money transfer using man-in-the-browser (MITB) attacks with malware in internet banking has become a social problem. This study focuses on transaction-tampering MITB attacks and considers countermeasures to combat them. The existing countermeasures require that another device, such as a token device, other than the device u...

In disaster management, timely delivery of the right information to the right people can improve outcomes and save lives. With fewer first responders available, volunteers play an increasing role in disaster assistance. A communication framework that brings victims, volunteers, and first responders together for timely disaster assistance can be inc...

This paper introduces a two-dimensional representation for trust values that uses two metrics: “trust” and “distrust.” With this representation, we can deal with such contradictory arguments as “The message is basically trustworthy but simultaneously not trustworthy.” Such situations can be caused when a message is consistent with other messages, b...

With the growth of image on the web, research on hashing which enables high-speed image retrieval has been actively studied. In recent years, various hashing methods based on deep neural networks have been proposed and achieved higher precision than the other hashing methods. In these methods, multiple losses for hash codes and the parameters of ne...

With the growth of image on the web, research on hashing which enables high-speed image retrieval has been actively studied. In recent years, various hashing methods based on deep neural networks have been proposed and achieved higher precision than the other hashing methods. In these methods, multiple losses for hash codes and the parameters of ne...

Cancelable biometric schemes have been widely studied to protect templates in biometric authentication over networks. These schemes transform biometric features, and perform pattern matching without restoring the original features. Although they strongly prevent the leakage of the original features, the response time can be very long in a large-sca...

中央集権的組織に依存せずに仮想通貨やスマートコントラクトを実現する分散台帳技術が近年注目されている．分散台帳技術が今後安全かつ信頼できる社会基盤として普及するためには，取引の健全性，クレデンシャルの非中央集権的管理，承認権限の独占困難性，という3つの特徴を備えることが必要となる．しかしながら，現実に運用されている分散台帳ではこれらの特徴すべてを備えるものはなく，安全かつ信頼できる台帳管理を行うことは依然として困難であった．そこで我々は，これらの特徴をすべて備える基盤として，ユーザと不可分な秘密情報をトラストアンカとしてトランザクションの生成・承認を行うHuman Trusted Distributed Ledger（HTDL）を提案し，その要件を整理した．さらに，生体情報を秘密鍵として用いて...

Biometric information can be easily leaked and/or copied. Therefore, the biometric information used for biometric authentication should be kept secure. To cope with this issue, we have proposed a user authentication system using a human reflex response. It is assumed that even if people know someone’s reflex characteristics, it is difficult to impe...

When faced with cross-site scripting (XSS) attacks, it is difficult to counter all malicious inputs such that they are rendered completely harmless. In such situations, the introduction of a whitelist-based XSS countermeasure is considered to be an effective and robust approach. However, as the behavior of current web applications is complex, it is...

Recently, biometric authentication has been applied to not only sensitive services such as in emigration/ immigration inspection systems and ATMs but also nonsensitive services such as entry/exit management systems of theme parks or coin lockers. However, biometric authentication requirements for sensitive and nonsensitive services differ; each ser...

A cancelable biometric scheme called correlation-invariant random filtering (CIRF) is known as a promising template protection scheme. This scheme transforms a biometric feature represented as an image via the 2D number theoretic transform (NTT) and random filtering. CIRF has perfect secrecy in that the transformed feature leaks no information abou...

A cancelable biometric scheme called correlation-invariant random filtering (CIRF) is known as a promising template protection scheme. This scheme transforms a biometric feature represented as an image via the 2D number theoretic transform (NTT) and random filtering. CIRF has perfect secrecy in that the transformed feature leaks no information abou...

In a blockchain system, a blockchain transaction is protected against forgery by adding a digital signature. By digital signature verification, we can confirm that a creator of a transaction has a correct private key. However, in some critical fields, we need to prove that a creator of a transaction is a proper user. In such a case, the conventiona...

Anonymization is one of the major processes to protect location-based services (LBS) from privacy leakage. However, there are many discussions about de-anonymization attacks to LBS and whether anonymization processing is a sufficient countermeasure for privacy leakage. In this paper, we proposed a novel method to reconstruct the location of user co...

With the rapid spread of smart devices, such as smartphones and tablet PCs, user authentication is becoming increasingly important because various kinds of data concerning user privacy are processed within them. At present, in the case of smart devices, password-based authentication is frequently used; however, biometric authentication has attracte...

We analyze the theoretical vulnerability of maximum a poste-riori(MAP) speaker adaptation, which is widely used in practical speaker recognition systems. First, we proved that there exist a set of feature vectors, what are called wolves, which can impersonate almost all the registered speakers with probability asymptotically close to 1 with at most...

Presentation attack against biometric authentication systems is getting attention by many researchers since the seminal paper[4]. In this paper, we consider a new kind of presentation attack called wolf attack. Wolf attack does not require biometric feature from a victim which the other presentation attacks does but synthesizes biometric feature fr...

Biometric cryptosystems have been widely studied in the literature to protect biometric templates. To ensure sufficient security of the biometric cryptosystem against the offline brute-force attack (also called the FAR attack), it is critical to reduce FAR of the system. One of the most effective approaches to improve the accuracy is multibiometric...

Impersonation by impostors is one of the representative security issues on biometric authentication system. A wolf attack is an attack on biometrics system using a wolf that can be falsely accepted as a match with multiple templates. False acceptance rate (FAR) which has been a conventional standard measure to quantify the average error rates of de...

Recently, with the rapid spread of smartphones and tablet PCs, biometric authentication using handwritten signatures has been attracting much interest. However, little research has been reported on biometric template protection, which is indispensable for protecting one's privacy, especially when those devices are used in a network environment. The...

Biometric template protection techniques are able to provide a solution for vulnerability which compromises biometric template. These are expected to be useful for remote biometric authentication. However, there were no supported standard evaluation method, thus making these technology not fully trusted. The draft Recommendation ITU-T X.gep: 'A gui...

Existing studies on the security of biometric authentication with template protection have considered the adversaries who obtain only protected templates. Since biometric authentication systems transmit data other than the protected templates, we need to consider how to secure biometric authentication systems against adversaries with those data. In...

Impersonation attack is one of the major security issues of biometric authentication systems. Wolf attacks use a biometric sample such that the similarities between this sample and a number of templates are resulting in high false matches with these templates. In the conventional evaluation with the wolf attack probability (WAP), wolf attacks took...

Biometric authentication based on template protection has attracted attention in the past decade. In the discussion on the security of these systems, however, the content of biometric information is assumed to be sufficiently large, and real conditions of biometric features have not yet been reflected. This paper focuses on a biometric cryptosystem...

In the present template protected biometric authentication method, which stores a user's biometric information in a template, there is a leakage risk of biometric information when the template is leaked. Therefore, we propose a new biometric cryptosystem which stores and restores a user's pre-set secret information and authenticate the user by usin...

We propose the minimum distance entropy (MDE) as a metric of biometric information content. The MDE is the probability that two biometric samples correspond exactly expressed in information content and can be calculated through the experiment for interpersonal matching using a set of biometric samples. This metric makes it possible for certain biom...

Biometric person authentication has been attracting considerable attention in recent years. Conventional biometric person authentication systems, however, simply store each user's template as-is on the system. If registered templates are not properly protected, the risk arises of template leakage to a third party and impersonation using biometric d...

Biometric person authentication has been attracting considerable attention in recent years. Conventional biometric person authentication systems, however, simply store each user's template as-is on the system. If registered templates are not properly protected, the risk arises of template leakage to a third party and impersonation using biometric d...