驗證碼為提高網路資訊安全的普遍形式,目前驗證碼多為透過靜態的文 字變形扭曲並結合背景圖像的機制,但逐漸受到光學文字辨識技術 (Optical Character Recognition, 簡稱OCR)的演進而遭到破解的威脅且缺乏安全性。因 此,本文以完形心理學(Gestalt Psychology)提出的封閉法則(Law Of Closure) 做為動態影像驗證碼的設計概念,將文字轉為動態圖形產生動態影像驗證碼。由於本驗證碼利用封閉法則的主觀輪廓概念進行設計,當使用各式形狀圖案與文字字元圖層相互重疊所產生出的不完整破碎文字圖形,人類大腦在 視覺感知中仍可以對不完整的圖形會自動產生無中生有,並將其缺口填補完整的錯視現象。本文提出的驗證碼,係將文字轉化為圖形再與幾合形狀產生差集並配合隨機律動結合雜訊背景層,以期增加動態影像驗證碼安全性。由於人類特有的錯視知覺現象,只有人類較能輕易的識別驗證碼內容,反之能 降低被惡意程式辨識與破解的機會。
CAPTCHA is a general form to improve the security of internet. The normally type of current CAPTCHA is combined with distorted words and noise background images. While the evolution of Optical Character Recognition there are more and more security threats of CAPTCHA. In this paper, we propose a Dynamic Image CAPTCHA. Our research use the Law Of Closure, which is a part of Gestalt laws of organization based on Gestalt psychology , as the design concept. Dynamic Image CAPTCHA uses incomplete character and shape to complement transform graphics, dynamic graphic layer and noise background layer to display a dynamic incomplete CAPTCHA image Based on human visual illusion, the dynamic incomplete images will seems to automatically generate to a completed image. The dynamic CAPTCHA which we proposed to improve the security is building by dynamic motion images and noise layer. Due to the illusion visual perception is especially for human, so that malware cannot crake the CAPTCHA easily. In other words, only human can recognize this dynamic CAPTCHA more effortlessly.