Personal Entropy from Graphical Passwords: Methods for Quantification and Practical Key Generation

Masato AKAO; Shinji YAMANAKA; Goichiro HANAOKA; Hideki IMAI

Personal Entropy from Graphical Passwords: Methods for Quantification and Practical Key Generation

Masato AKAO, Shinji YAMANAKA, Goichiro HANAOKA, Hideki IMAI

Full Text Views

0

Cite this

Summary :

In many cryptosystems incorporating human beings, the users' limited memories and their indifference to keeping the systems secure may cause some severe vulnerability of the whole systems. Thus we need more studies on personal entropy, from an information theoretical point of view, to capture the characteristics of human beings as special information sources for cryptosystems. In this paper, we discuss and analyze the use of personal entropy for generating cryptographic keys. In such a case, it is crucially important to precisely evaluate the amount of personal entropy that indicates the actual key length. We propose an advanced key generation scheme based on the conventional graphical passwords proposed in [12]. We improve them to make the most of the secret information extracted in one drawing, i.e., we incorporate the on-line pen pressure and pen inclination information in addition to utilize more secret information. We call the scheme dynamic graphical passwords, and propose a practical construction of them. We also show a precise way of quantifying their entropy, and finally, as an experimental result, we can generate a key of over 110-bit long, using the data of a single drawing. When quantifying their entropy, we need to precisely evaluate the entropy of graphical passwords as well as that of the on-line information of pen movements. We need to precisely evaluate the entropy of graphical passwords by considering the users' biased choices of their graphical passwords. It is expected that they tend to choose their passwords that are memorable as easily as possible, thus we quantify the burden of memorizing each graphical password by the length of its description using a special language based on [12]. We improve the approach in [12] by more directly reflecting how easily each graphical password can be memorized.

Publication: IEICE TRANSACTIONS on Fundamentals Vol.E87-A No.10 pp.2543-2554

Publication Date: 2004/10/01

Publicized

Online ISSN

DOI

Type of Manuscript: Special Section PAPER (Special Section on Information Theory and Its Applications)

Category: Cryptography and Information Security

Cite this

Copy

Masato AKAO, Shinji YAMANAKA, Goichiro HANAOKA, Hideki IMAI, "Personal Entropy from Graphical Passwords: Methods for Quantification and Practical Key Generation" in IEICE TRANSACTIONS on Fundamentals, vol. E87-A, no. 10, pp. 2543-2554, October 2004, doi: .
Abstract: In many cryptosystems incorporating human beings, the users' limited memories and their indifference to keeping the systems secure may cause some severe vulnerability of the whole systems. Thus we need more studies on personal entropy, from an information theoretical point of view, to capture the characteristics of human beings as special information sources for cryptosystems. In this paper, we discuss and analyze the use of personal entropy for generating cryptographic keys. In such a case, it is crucially important to precisely evaluate the amount of personal entropy that indicates the actual key length. We propose an advanced key generation scheme based on the conventional graphical passwords proposed in [12]. We improve them to make the most of the secret information extracted in one drawing, i.e., we incorporate the on-line pen pressure and pen inclination information in addition to utilize more secret information. We call the scheme dynamic graphical passwords, and propose a practical construction of them. We also show a precise way of quantifying their entropy, and finally, as an experimental result, we can generate a key of over 110-bit long, using the data of a single drawing. When quantifying their entropy, we need to precisely evaluate the entropy of graphical passwords as well as that of the on-line information of pen movements. We need to precisely evaluate the entropy of graphical passwords by considering the users' biased choices of their graphical passwords. It is expected that they tend to choose their passwords that are memorable as easily as possible, thus we quantify the burden of memorizing each graphical password by the length of its description using a special language based on [12]. We improve the approach in [12] by more directly reflecting how easily each graphical password can be memorized.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/e87-a_10_2543/_p

Copy

@ARTICLE{e87-a_10_2543,
author={Masato AKAO, Shinji YAMANAKA, Goichiro HANAOKA, Hideki IMAI, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Personal Entropy from Graphical Passwords: Methods for Quantification and Practical Key Generation},
year={2004},
volume={E87-A},
number={10},
pages={2543-2554},
abstract={In many cryptosystems incorporating human beings, the users' limited memories and their indifference to keeping the systems secure may cause some severe vulnerability of the whole systems. Thus we need more studies on personal entropy, from an information theoretical point of view, to capture the characteristics of human beings as special information sources for cryptosystems. In this paper, we discuss and analyze the use of personal entropy for generating cryptographic keys. In such a case, it is crucially important to precisely evaluate the amount of personal entropy that indicates the actual key length. We propose an advanced key generation scheme based on the conventional graphical passwords proposed in [12]. We improve them to make the most of the secret information extracted in one drawing, i.e., we incorporate the on-line pen pressure and pen inclination information in addition to utilize more secret information. We call the scheme dynamic graphical passwords, and propose a practical construction of them. We also show a precise way of quantifying their entropy, and finally, as an experimental result, we can generate a key of over 110-bit long, using the data of a single drawing. When quantifying their entropy, we need to precisely evaluate the entropy of graphical passwords as well as that of the on-line information of pen movements. We need to precisely evaluate the entropy of graphical passwords by considering the users' biased choices of their graphical passwords. It is expected that they tend to choose their passwords that are memorable as easily as possible, thus we quantify the burden of memorizing each graphical password by the length of its description using a special language based on [12]. We improve the approach in [12] by more directly reflecting how easily each graphical password can be memorized.},
keywords={},
doi={},
ISSN={},
month={October},}

Copy

TY - JOUR
TI - Personal Entropy from Graphical Passwords: Methods for Quantification and Practical Key Generation
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 2543
EP - 2554
AU - Masato AKAO
AU - Shinji YAMANAKA
AU - Goichiro HANAOKA
AU - Hideki IMAI
PY - 2004
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E87-A
IS - 10
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - October 2004
AB - In many cryptosystems incorporating human beings, the users' limited memories and their indifference to keeping the systems secure may cause some severe vulnerability of the whole systems. Thus we need more studies on personal entropy, from an information theoretical point of view, to capture the characteristics of human beings as special information sources for cryptosystems. In this paper, we discuss and analyze the use of personal entropy for generating cryptographic keys. In such a case, it is crucially important to precisely evaluate the amount of personal entropy that indicates the actual key length. We propose an advanced key generation scheme based on the conventional graphical passwords proposed in [12]. We improve them to make the most of the secret information extracted in one drawing, i.e., we incorporate the on-line pen pressure and pen inclination information in addition to utilize more secret information. We call the scheme dynamic graphical passwords, and propose a practical construction of them. We also show a precise way of quantifying their entropy, and finally, as an experimental result, we can generate a key of over 110-bit long, using the data of a single drawing. When quantifying their entropy, we need to precisely evaluate the entropy of graphical passwords as well as that of the on-line information of pen movements. We need to precisely evaluate the entropy of graphical passwords by considering the users' biased choices of their graphical passwords. It is expected that they tend to choose their passwords that are memorable as easily as possible, thus we quantify the burden of memorizing each graphical password by the length of its description using a special language based on [12]. We improve the approach in [12] by more directly reflecting how easily each graphical password can be memorized.
ER -