QUIZ 1 Font selection as The end user was presented with a set of images and asked to create the user-perceived concept of the class of pleasing images. The concept represents the rule by which a new image can be classified as belonging to the defined class. As one of such cases, consider the task of finding an eye-pleasing font for a concrete user. The user is asked to classify different examples of fonts for GUI. The user marks the examples as positive or negative. Thus a concept of what the user likes will be formed. Now another set of fonts should be used (for example, when some new mathematical or language specific characters are required). Which of the new fonts will be favored by the user? To keep this task simple, let's say that the characters have only one characteristic significant for the user the density of the font measured as the ratio of black pixels to their total number in the rasterized font's mage The concept of the eye-pleasing fonts for an individual user is presented as set of the following 16 examples one half of which the user perceived positively and the other half negatively: Positive perception Font density: 0.7, 0.6, 0.3, 0.8, 0.4, 0.9, 0.5, 0.8 Negative perception Font density: 0.1, 0.4, 0.2, 0.5, 0.3, 0.6, 0.3, 0.3 Now a new font was selected with five additional characters having the following densities 0.4, 0.6, 0.7, 0.5,0.8 Will this font be well received by the user? Note Consider this task as the classification problem. In this case, we deal with two image clusters of positive and negative perceptions, accordingly. When a new image is presented, the task is to find out which cluster this image, most likely, belongs to Refer to the paper as posted CueFlik: Interactive Concept Learning in Image Search James Fogarty, Desney S Tan, Ashish Kapoor, Simon Winder. Proceedings of ACM CHI 2008 Here are a few relevant excerpts Popular image search engines have begun to provide tags based on simple characteristics of images (such as tags for black and white images or images that contain a face), but such approaches are limited by the fact that it is unclear what tags end users want to be able to use in examining image search results. CueFlik is an image search application that allows end users to quickly create (and reuse) their own rules for re-ranking images based on their visual characteristics. QUIZ 1 Font selection as The end user was presented with a set of images and asked to create the user-perceived concept of the class of pleasing images. The concept represents the rule by which a new image can be classified as belonging to the defined class. As one of such cases, consider the task of finding an eye-pleasing font for a concrete user. The user is asked to classify different examples of fonts for GUI. The user marks the examples as positive or negative. Thus a concept of what the user likes will be formed. Now another set of fonts should be used (for example, when some new mathematical or language specific characters are required). Which of the new fonts will be favored by the user? To keep this task simple, let's say that the characters have only one characteristic significant for the user the density of the font measured as the ratio of black pixels to their total number in the rasterized font's mage The concept of the eye-pleasing fonts for an individual user is presented as set of the following 16 examples one half of which the user perceived positively and the other half negatively: Positive perception Font density: 0.7, 0.6, 0.3, 0.8, 0.4, 0.9, 0.5, 0.8 Negative perception Font density: 0.1, 0.4, 0.2, 0.5, 0.3, 0.6, 0.3, 0.3 Now a new font was selected with five additional characters having the following densities 0.4, 0.6, 0.7, 0.5,0.8 Will this font be well received by the user? Note Consider this task as the classification problem. In this case, we deal with two image clusters of positive and negative perceptions, accordingly. When a new image is presented, the task is to find out which cluster this image, most likely, belongs to Refer to the paper as posted CueFlik: Interactive Concept Learning in Image Search James Fogarty, Desney S Tan, Ashish Kapoor, Simon Winder. Proceedings of ACM CHI 2008 Here are a few relevant excerpts Popular image search engines have begun to provide tags based on simple characteristics of images (such as tags for black and white images or images that contain a face), but such approaches are limited by the fact that it is unclear what tags end users want to be able to use in examining image search results. CueFlik is an image search application that allows end users to quickly create (and reuse) their own rules for re-ranking images based on their visual characteristics