Image-based CAPTCHA generation system
In a system and method for the generation of attack-resistant, user-friendly, image-based CAPTCHAs (Completely Automated Public test to Tell Computers and Humans Apart), controlled distortions are applied to randomly chosen images and presented to a user for annotation from a given list of words. An image is presented that contains multiple connected but independent images with the borders between them distorted or otherwise visually obfuscated in a way that a computer cannot distinguish the borders and a user selects near the center of one of the images. The distortions are performed in a way that satisfies the incongruous requirements of low perceptual degradation and high resistance to attack by content-based image retrieval systems. Word choices are carefully generated to avoid ambiguity as well as to avoid attacks based on the choices themselves.
1. A method of generating a Completely Automated Public test to Tell Computers and Humans Apart (CAPTCHA), comprising the steps of:
storing a database of images and annotations corresponding to the images, each image depicting a concept or object intended for human recognition;
distorting at least one of the images, and presenting the distorted image on a computer display along with a separate list of words, one of which best annotates the distorted image; and assuming that the user of the computer is a human as opposed to a machine if the correct word is selected to annotate the distorted image.
2. The method of claim 1 , wherein the image is distorted using one or more of the following:
dithering, partitioning, quantization, noise addition, color re-mapping, and selective cut-and-resize.
3. The method of claim 1 , further including the steps of:
displaying a plurality of different images prior to the step of displaying a distorted image to be annotated;
selecting one of the images from the plurality; and
displaying the distorted image and the separate list of words, one of which best annotates the selected image.
4. The method of claim 3 ,
wherein the plurality of images are separated by borders which are distorted or otherwise visually obfuscated.
5. The method of claim 1 , wherein the words are generated to avoid ambiguity with respect to the selected image.
6. A method of generating a Completely Automated Public test to Tell Computers and Humans Apart (CAPTCHA), comprising the steps of:
storing a database of images and annotations corresponding to the images, each image depicting a concept or object intended for human recognition;
presenting a composite image on a computer display, the composite image including plurality of different images randomly selected from the database;
selecting one of the images from the composite image;
distorting the selected image and presenting the distorted image along with a separate list of words, one of which best annotates the distorted image; and
assuming that the recipient of the image is a human as opposed to a machine if the correct word is selected to annotate the distorted image.
7. The method of claim 6 , wherein the image is distorted using one or more of the following:
dithering, partitioning, quantization, noise addition, color re-mapping, and selective cut-and-resize.
8. The method of claim 6 , wherein the plurality of images are separated by borders which are distorted or otherwise visually obfuscated.
9. The method of claim 6 , wherein the words are generated to avoid ambiguity.
10. The method of claim 6 , wherein the image is distorted using content-based image retrieval and annotation.
11. The method of claim 6 , further including the steps of:
selecting one of the images from the composite image by clicking in an area at or near the geometric center of one of the images in the composite image; and
wherein the process is repeated if the clicking does not occur at or near the geometric center of one of the images.
12. A Completely Automated Public test to Tell Computers and Humans Apart (CAPTCHA) system, comprising:
a database for storing annotated images, each image depicting a concept or object intended for human recognition;
a processor for distorting the images;
a display for presenting a distorted image to a user along with a separate list of annotations to choose from; and
an input allowing a user to select the annotation that best corresponds to the distorted image.
13. The CAPTCHA system of claim 12 , wherein the processor is operative to display a composite image containing a plurality of images.
14. The CAPTCHA system of claim 12 , wherein the input allows a user to select one of the images from the composite image prior to annotation.
15. The CAPTCHA system of claim 12 , wherein the processor is operative to:
present a composite image including a plurality of different images separated by borders which are distorted or otherwise visually obfuscated.