CAPTCHAs are commonly used as gatekeepers to protect online services from automated bots. Previous CAPTCHA evaluations have solely focussed on whether CAPTCHA challenges are unbreakable by bots but solvable by humans and have ignored a fundamental question: what effect do these gatekeepers have on their human users? Our study proposes a more realistic approach that complements existing CAPTCHA evaluations by measuring CAPTCHA’s effect on the main task that the user wants to complete. Through our experimental setup, we show that failing to correctly answer CAPTCHAs can turn away a significant percentage of legitimate, human users. We also show that these failures have temporary knock-on effects on the quality of tasks that users later perform. Methodologically, our study also reveals limitations of current approaches to batch evaluations of CAPTCHAs that do not accurately capture the effects that ordering and variance in difficulty have on users’ CAPTCHA-solving performance.

This paper reconsiders the use of MOS naturalness as an instrument for measuring the quality (vs. intelligibility) of speech. We reconsider an earlier proposed alternative, the paired comparison or “AB” test, and present new empirical evidence that this is indeed a better method for evaluating TTS quality. Using this, we evaluate three older TTS systems along with a recent deep-learning approach against native North-American and Indian speech and show that, in fact, TTS had already crossed the threshold of human-like speech synthesis some time ago. This suggests that a systematic reappraisal of the concept of abstract “naturalness” of speech is in order.

One of the limitations of wildcard rules in Software Defined Networks, such as OpenFlow, is losing visibility. FleXam is a flexible sampling extension for OpenFlow that allows the controller to define which packets should be sampled, what parts of each packet should be selected, and where they should be sent. Here, we present an interactive demo showing how FleXam enables the controller to dynamically adjust sampling rates and change the sampling scheme to optimally keep up with a sampling budget in the context of a traffic statistics collection application.

Speech certainly has advantages as an input modality for smartphone applications, especially in scenarios where using touch or keyboard entry is difficult, on increasingly miniaturized devices where useable keyboards are difficult to accommodate, or in scenarios where only small amounts of text need to be input, such as when entering SMS texts or responding to a CAPTCHA challenge. In this paper, we propose two new alternative ways to design CAPTCHAs in which the user says the answer instead of typing it with (a) output stimuli provided visually (SeeSay) or (b) auditorily (HearSay). Our user study results show that SeeSay CAPTCHA requires less time to be solved and users prefer it over current text-based CAPTCHA methods.

Speech authentication protocols should have a challenge/response feature to be protected against replay attacks. As a result, they need to verify whether the user responded to an interactive prompt. However, it is usually assumed that the user will provide their answer perfectly. In this paper, we report on an ecologically valid user study that we conducted to test this assumption. Our results show that 40% of user answers are imperfect, even in a task as simple as sentence repetition. Error analysis reveals that 60% of the imperfect answers contain small errors that should be deemed acceptable, which increases the total acceptance rate of this task to 84%. We also tested a forced alignment algorithm as a means of verifying answers automatically.

Verifying that a live human is interacting with an automated speech based system is needed in some applications such as biometric authentication. In this paper, we present a method to verify that the user is human. Simply stated, our method asks the user to repeat a sentence. The reply is analyzed to verify that it is the requested sentence and said by a human, not a speech synthesis system. Our method is taking advantage of both speech synthesizer and speech recognizer limitations to detect computer programs, which is new, and potentially more accessible, way to develop CAPTCHA systems. Using an acoustic model trained on voices of over 1000 users, our system can verify the user’s answer with 98% accuracy and with 80% success in distinguishing humans from computers. Index Terms: Accessibility, CAPTCHA, Speech Recognition, Speech Synthesis

Embedding a secret message into a cover media without attracting any attention, known as steganography, is one of the methods used for hidden communication purposes. One of the cover media that can be used for steganography is speech. In this study, the authors propose a new steganography method in speech signals. In this method, the silence intervals of speech are found and the length (number of samples) of these intervals is changed to hide information. The main feature of our method is robustness to MPEG-1 layer III (MP3) compression. This method can hide information in a speech stream with very low processing time which makes it a real-time steganography method. The hiding capacity of our method is comparable with other MP3 resistance methods and the listening tests show that the degradation in speech quality is not annoying. Additionally, the effect of our method on chaotic features is negligible, so it is difficult to detect our method with chaotic-based steganalysis methods.

Multimedia files including audio, music and movies are important part of resources available on the Internet. With the expansion of personal computers and other digital devices such as digital cameras, ordinary people are also creating different multimedia files. Hence, there are huge amount of multimedia files in the Internet. Considering the volume of files available, we need automatic methods for indexing, classification and retrieval of multimedia files. But, there are a number of problems in processing multimedia files. Two important issues are the large size of multimedia files and different lengths of different files.In this PhD project, we want to develop a solution for these two problems. We design a signature creation algorithm that creates fixed length presentation of multimedia files. This signature can be used to measure similarity of different files and enables us to index and search among multimedia files. Currently we apply our method on music files and it performs well for tasks such as artist identification and genre classification.

Covert exchange of information has been long a significant and highlighted affair. Numerous methods have been devised for the matter of which steganography is a popular one. Already applied in the past, the method has been massive used following the invention of computer. This paper discusses a new application of steganography in textiles in which information is hidden in the images on the clothes. In this way, information is first hidden by computer in images by a special method. Then the provided images are printed on clothes. The related information is extracted after scanning the images by relevant programs.

Sending information secretly and communicating covertly have been of great interest for ages. On the other hand, text documents have been widely used and consequently various methods for hiding information in texts (text steganography) have been developed so far. In this paper a new method is proposed to hide information in digital Persian and Arabic Unicode texts.In Persian and Arabic, each letter can have four different shapes regarding to its position in the word. In this method by using this feature of Persian and Arabic languages and the way which documents are saved in the Unicode Standard, the information is hidden in Unicode text documents. This method has a good hiding capacity because it hides one bit in each word. Also this method does not make any apparent changes in the original text and have a perfect perceptual transparency.

Nowadays the Internet users are from different ages and groups. Disabled people are a group of the Internet users. Some Web sites are especially created for these people. One group of the disabled people are deaf persons. They have a special talking language which is named sign language. Here we present a method to encourage them, esp. children, to learn the sign language. In this method, when a deaf person wants to enter a Web site which is created for deaf persons, a word is shown as a movie using a sign language. The user should recognize the word and select it from a list. If the user understands the sign language and recognizes the word, he/she can enter the website. This project has been implemented by PHP scripting language.

Steganography is the art of hiding information in a cover media without attracting attention. One of the cover media which can be used for steganography is speech. In this paper, we propose a new speech steganography in wavelet domain. In this method, lifting scheme is used to create perfect reconstruction Int2Int wavelets. The data is hidden in some of the Least Significant Bits (LSB) of detail wavelet coefficients. The LSB bits for hiding are selected with a new adaptive algorithm. This algorithm does not hide information in silent parts, so there is no need for silent detection algorithms. This method has zero error in hiding/unhiding process, while normal wavelet domain LSB has about 0.2 % error in equal hiding capacity. This method is a high capacity steganography method which can hide information up to 20% of the input speech. The Signal-to- Noise Ratio (SNR) and listening tests show that the stegano audio is imperceptible from original audio.