The image below is from a language learning app in which users guess which image from a selection of four matches a recorded audio caption. All of the images and captions are crowd sourced. Currently the app makes a random selection about which images and captions to serve. But I'd like to introduce a feature to allow users to select a category to narrow the domain of possible images shown. For example, the user might select "Transportation" or "Fruits" to practice only the vocabulary concepts from that particular category. Furthermore because the database is constantly expanding as users contribute more content I'd like to indicate for each category how many images are available. For example instead of "Transportation" it might say "Transportation (65)" indicating that there are 65 images and captions available in that category to practice with.
Given the screen below, how would you suggest introducing a UI element to enable the toggling of categories?