@techreport{oai:ipsj.ixsq.nii.ac.jp:00218231,
 author = {花田, 雄一 and 内田, 秀継 and 長村, 一樹 and 舟橋, 涼一 and 松山, 佳彦 and 安部, 登樹 and 山田, 茂史 and Yuichi, Hanada and Hidetsugu, Uchida and Kazuki, Osamura and Ryoichi, Funabashi and Yoshihiro, Matsuyama and Narishige, Abe and Shigefumi, Yamada},
 issue = {8},
 month = {Jun},
 note = {高い利便性と高いセキュリティを併せ持つ生体認証の活用が進んでいる．特に，店舗での決済やオフィスの入退場管理として，ID 情報を入力せず生体情報だけで本人確認する手ぶらでの認証が注目されており，ユーザ数の増加が見込まれている．しかしながら，手ぶらでの認証では，全登録者の生体情報との突合が必要なため，登録者数が増加すると認証処理時間も増加し，サーバ負荷やシステムコストの肥大化につながる．そこで，認証処理の高速化，負荷低減のために，ある認証ポイントをよく利用する人の登録生体情報を事前にキャッシュしてそれと認証する方式（キャッシュ認証）があるが，素朴な手法ではある認証ポイントを最初に訪れるユーザは本効果の対象外となり，キャッシュヒット率向上に限界があった．本稿では，認証サービス利用履歴に基づいて，複数の認証ポイントをクラスタリングしておくことで，初回ユーザであってもキャッシュ認証の効果を享受できる手法を提案する．評価実験ではユーザの位置情報データを来店と見立てた来店データを使用して従来手法との比較評価によりアルゴリズムの有効性を示した．, The use of biometric authentication, which has both high convenience and high security, is advancing. In particular, empty-handed authentication, in which identity verification is performed only with biometric information without entering ID information, is attracting attention as payment at stores and entrance / exit management of offices, and the number of users is expected to increase. However, since empty-handed authentication requires matching with the biometric information of all registrants, as the number of users increases, the authentication processing time also increases, leading to an increase in server load and system cost. Therefore, in order to speed up the authentication process and reduce the load, there is a method (cache authentication) that caches the registered biometric information of a person who frequently uses a certain authentication point in advance and authenticates it, but it is a simple method. The user who visits the point first is not subject to this effect, and there is a limit to the improvement of the cache hit rate. In this paper, we propose a method that allows even first-time users to enjoy the effects of cache authentication by clustering multiple authentication points based on the authentication service usage history. In the evaluation experiment, the effectiveness of the algorithm was shown by comparative evaluation with the conventional method using the store visit data that regards the user's location information data as a store visit.},
 title = {認証サービス履歴に基づいた照合対象者選択技術の検討},
 year = {2022}
}