CyberSiara™  |  Advanced Security

A computer can set up thousands of fake email accounts in just seconds, or generate whole new fake identities causing chaos to many businesses. If you look at the variety of computer-based attacks now being deployed, it's challenging to secure a site. To aid businesses in this daily threat, CybersiaraTM developed our unique human verification tool, called SiaraShieldTM.

SiaraShield for security

SiaraShieldTM has been independently tested and evaluated under advanced laboratory conditions as part of an extensive program of intensive security and machine learning tests. The results confirm that SiaraShieldTM is robust against the most sophisticated optical character recognition (OCR) software and advanced image recognition software on the market.

SiaraShield for security

SiaraShieldTM was also tested using advanced artificial intelligence techniques and deep learning mechanisms in one of the most advanced information security laboratories in the UK*. The results were simply outstanding. Not a single one of the current image recognition and AI mechanisms were able to recognise and bypass SiaraShieldTM.

SiaraShield for security

Using our ingenious technology, we have managed to create a reliable and robust solution to protect online businesses of all sizes. SiaraShieldTM is the most reliable human verification tool on the market, and it is also one of the fastest and most convenient for businesses to deploy and website users to authenticate.

Current Security challenges

The machine learning algorithms can use the several methods and tools to defeat Captchas: Optical Character Recognition, (OCR) Segmentation, Colour Matching Technique, Brute Force Attack (Dictionary Attack), and the highly complex Frames Aggregation & Deep Learning Techniques.

Bots are becoming increasingly successful at deciphering and recognising the current verification tests and Captchas. According to scientific research, nearly all of the current text-based and image-based tests on the market are vulnerable against these recognition methods.

Websites are being relentlessly targeted by automated computer attacks. Businesses are often unprepared and do not have the resources to manage the disruption this causes. Despite putting various protective mechanisms in place on their websites, businesses are still being successfully penetrated by malicious bots.

The results of our advanced security experiments confirm that our technology is a breakthrough against these bot attacks. SiaraShieldTM is 100% secure* against current character recognition and advanced techniques. Businesses can now enjoy peace of mind and confidence that their websites are well protected against ever-increasing bot attacks.

SiaraShield for challange

Brute Force Attack (or Dictionary Attack)

Brute Force (also known as Dictionary) is the most simplistic technique used to solved existing Human Verification Tools. Utilizing a vast database, bots randomly apply previously solved tests until a match is found. With the ability to process thousands of attempts in a short amount of time, this technique can be very successful with simpler verification tests.

SiaraShieldTM protects against Brute Force by generating a new alphanumeric code after every incorrect input. There are 7,962,624 different combination and generating a new code after every incorrect answer means bots have a near 0% chance for a Brute Force scheme to defeat SiaraShieldTM.

Human Verification Tests have evolved to protect against Brute Force techniques, by layering distorted characters over colourful patterns and random lines. Bots have increased in their complexity by employing Advanced Optical Character Recognition (OCR) to solve verification tests, often at a better rate than humans. Using OCR, bots can easily recognise and decipher characters and images, and the technology continues to improve. The two most common applications of OCR are Segmentation and Colour Matching, which are often used in conjunction to solve tests:

Optical Character Recognition (OCR)

Advanced OCR software can easily recognise and decipher the characters by using different techniques as follow

Segmentation

The other method that OCR software uses to recognise text, is by segmenting and separating the characters and isolate them in order to analyse them. Then the software would be able to use another mechanism to map the characters with its database to find the closest match. The example below is showing an example of one of the current captcha models on the market. As it can be seen from the demonstration, it is very easy to break this type of captchas based on the colour difference between the object and the background noise.

SiaraShieldTM is armed with advanced algorithm to display only 'partial information' of the characters based on individual frames and wrap the partial information with background noise. Therefore, by analysing every single frame, no useful information about the object can be retrieved. Thus, a computer recognition programme would not be able to identify the characters and letters in order to separate them.

CyberSiara Ocr
Current steps involved in recognising and deciphering a captcha image by OCR software.

Colour Matching Technique

This technique analyses the image for differences in using colours or patterns. Once the characters are separated from the background, a common OCR or Segmentation process is applied to easily decipher the code.

This diagram shows how a popular type of Human Verification Test is defeated using the difference in colour between the object and the background noise. The characters are simply separated from the background noise, allowing a bot to match the well-defined characters to those in a database. With this process, a bot has the ability to solve thousands of tests of the same type, rendering them useless.

01 SiaraShield for security captcha

An Example of a text-based CAPTCHA model as it can be seen from the image, the text is printed in black against a colourful background.

02 SiaraShield for security captcha

The final output rendered image. The background is completely removed using "Colour Matching" techniques and the black text is only remaining in the image.

03 SiaraShield for security captcha

The figure shows after the background noise is removed and the text has been recognised using segmentation techniques.

SiaraShieldTM uses monochrome 1-bit pixels. Since the object and the background noise use only black or white, it is nearly impossible for bots to separate the object pixels from the background noise.

Frames Aggregation Technique

In order to increase site security, some verification tools are now based on video rather than a single still image. With a Frames Aggregation Technique, the video verification test is converted into a series of ordered frames. The bot compares pixels in the sequence to determine motion of the object.

01

SiaraShieldTM uses a smart noise filter to make the sequence of the frames unpredictable, making it impossible to order the frames for analysis of motion.

02

SiaraShieldTM superimposes frames based on human perception of density. The speed of the frames allows our eyes to see the hidden characters very easily. Even if a machine learning system superimposes the SiaraShieldTM frames, the result will be random noise and no image to analyze.

03

SiaraShieldTM is further enhanced by injecting random frames along the original frames on the sequence. These frames are displaying only briefly and invisible to the human eye, but are seen by bots, making the creation of an image to decipher even more difficult.