User:AalekhN/notes

Indexing Annotations Type Captcha

Following points suggested by community members , I have drafted few points which are mentioned below :

Modifying "annotations based questions", we can modify annotations based question to have ReCaptcha like solution, which would help us index images for Mediawiki site. An example for the question is shown here:

Now the only provided index for the images in the Wiki commons database is "Cat", while selecting the correct options for the image, correct answer ( Options ) to the question is: "Cat" and "Humans"

Utilizing Indexed System

These indexed photos can be provided as question to Visually Impaired but not completely blind user's . For Example: Image as options can be provided to Visually Impaired user's without manipulating them with any kind effect.

Advantage

These logical question are difficult for bots to solve since index's developed as an overtime Captcha Run are stored in Mediawiki Sites own database.

For blind and visually impaired users

Rather than developing a completely new type of captcha the approach could be to improve the interface of pre-existing Audio Captcha.The approaches to improve the captcha could be: 1) Assigning a key to make an audio playback. This audio playback for the same sound can be made for 3 times. After which the voice and the number spoken will change. 2) Assigning a key to make delay in the captcha voice for small duration of time. 3) Assigning a key to reload a particular number in the captcha, rather than reloading the whole captcha again a key could be assigned to reload a specific portion/number of the captcha. More information about the improvement of the captcha could be found here.

Selecting/Customizing categories for images to be displayed

An additional point to mention about customizing categories, there can be an array of categories similar to array of character in Questy Captcha, these categories can be customized by Wiki owner's. The use Categories will be to fetch images belonging to that category from wikidata. For example:let us define an array of categories as array('Human','Animals','Ocean','Mountain').Now the Captcha will fetch images for questions from these given categories only.

Advantage

• It would provide Wiki administrator with opportunity to customize the image,hence providing effective localization for the Captcha
• Enhances the Captcha with an robust feature of Questy Captcha ,alongwith adding more possible images for a given Captcha.

How to make these categories Unrelated

In order to make categories unrelated we can make super-set of the categories for example:-> there could be categories of artist,astronauts can be categorized under the Super Category of humans, similarly collection of an array can be made comprising of unrelated categories such as “people”, “animals”, “machines" etc...,moreover this array of unrelated categories can be modified by administrators of Wiki's to add more categories according to his need.

Captcha for Mobile Device

The Captcha for the mobile devices will be shown over Lightbox effect produced with the help of JQuery.

For "annotation type captcha" the design over lightbox would be like:

if the users clicks on the correct option next image to be verified will be shown as shown below:

If the user clicks on an wrong option all the image options of the captcha will be reloaded, and different images with options will be shown.

Moreover for "Odd one out" type question the layout of the question over the lightbox will be as shown:

Similar to annotation type captcha, if the user clicks on an wrong option all the image options of the captcha will be reloaded.

Annotations Based Captcha

This type of Captcha asks user to identify the objects in the image out of the given option's. The question for the given type of Captcha will be: the correct option's to the given type of question's are 'Cat' and 'Mountain' respectively. A demo representing raw working Prototype of this type of Captcha

is given here .

Advantage Over Bots  : Using image effects produced by ImageMagick Library we can make images non recognizable to the bots but on the 

same hand recognizable to the humans, for example applying photo booth and pencil sketch effect on the images produces images 

recognizable to the humans and non recognizable to bots, moreover number of option's provided makes it an advantage over current captcha 

system (based on probability of selecting the object).

The effect's used in the photographs can be used wisely in order to render human recognizable photographs also an an advantage images manipulated with different images can help us in prevention against the autonomous programs as mentioned in a research paper here .

Captcha based on Clipart

This type of Captcha asks user to click on the clipart asked. A sample question for the given type of Captcha is shown below:

In order to implement this type Captcha we can use an Open source PHP tool namely Visual Captcha. The translation feature to question can be implemented using Open Source Apertium Translation which provides support to wide number of major languages.A demo to the given type of Captcha is given here.

Advantage Over Bots  : According to visualcaptcha.net there hasn't been any report's of bots bypassing this captcha.

Ask User to click on the same image as provided

The Question for the given type of Captcha looks like:

the answer to given Captcha will be:

for better security issue I plan to provide 8 options for the Captcha rather than 4 options also we can provide multiple effects to an option for example: the above given file is produced by providing "pencil sketch+photo booth" effect to the below given image and is not recognizable to large image database's like google images and tineye api.

A demo representing raw working prototype of this Captcha is given here

Image Rotation Based Captcha

The following type of Captcha has been proposed by Google Research in an article here. Question of the given type of captcha will be to orient the images in an upside down position as shown here: we can rely selection made by the people overtime for the images which are uploaded in an improper orientation or are symmetric. For example: In the symmetric object (C) shown above the solution would be to rotate object by 90 degree in clockwise direction.