Study Of Chinese And English Text Steganography Based On Typos

Nowadays,information hiding technology plays an irreplaceable role in information security such as covert communication,copyright protection,source tracing and so on.At present,research on information hiding using the redundancy of human senses on carriers such as images,audio,video,and documents is very rich.However,due to the high abstraction and high semantic density of plain text,these carries are sensitive to modification and there is less redundancy for information hiding.Therefore,there are fewer plain text information hiding methods than those who use image,audio,video as carriers.But the problem of copyright protection,text leakage,and under-utilization of the plain texts on the Internet remains to be resolved.Input errors,also known as typos,usually occur in texts in the Internet such as social media.They also occur in texts of other types and languages at different frequencies.Therefore,this paper researches input errors and automatic correction,and finds that input error is a cross-lingual,widespread,and natural phenomenon that can be creatively integrated with information hiding.This paper first researches and classifies input errors in Chinese text and English text based on texts collected from social media and other environments and then designs a simple primary method to simulate input errors for text information hiding.With the above research and the study of text information hiding,this paper proposes a novel Chinese and English text information hiding method based on input errors using a natural language processing model named BERT.Due to the differences between Chinese and English,and the word-based embedding method's need of Chinese word segmentation,this paper proposes information embedding methods and blind extraction for Chinese text and English text,respectively.This method injects secret information into plain text by replacing some characters with generated typos who will not frustrate human understanding the text.Finally,the proposed method is analyzed by experiments to ensure its feasibility.This paper creatively proposed a novel information hiding method for Chinese texts and it has a low probability to modify the untouchable words.Also,this flexible method can be optimized by following up with related technologies.The method proposed in this paper can be applied to information security applications such as covert communication and digital watermarking.It provides a new method for hiding information in plain text,a new option for cross-media information hiding,and a new idea for other-lingual and cross-lingual information hiding.