Not Hot Dog
A fun app inspired by Silicon Valley's SeeFood app that uses ML to detect if food is a hot dog or not.
Well, if you’re a fan of Silicon Valley, you have definitely wanted to build Jian Yang’s breakthrough SeeFood app. If you don’t know what SeeFood is, it’s an app that tells you whether your food is a hot dog or not a hot dog. Pretty groundbreaking, right? 😂
It took me around two hours to build the complete application. The UI part was pretty simple: two screens, one to take a picture of the food and the other to detect if the food is a hot dog or not. The first screen has a live viewfinder and two buttons, one to capture the image from the viewfinder and the other to select an image from the gallery. For the live viewfinder, I used the camera plugin. Once an image is selected, the application moves to the second screen, where machine learning labels the image. For that, I used the mlkit plugin with Firebase’s label detection.
 |  |  |
The LabelDetector returned a list of VisionLabel objects, whose label property holds what the image contains. To find out if the image contained a hot dog, I checked whether the list held a VisionLabel labeled “Hot dog”, and after a bit of tinkering, the app was ready.
To make my app different from Jian Yang’s, I also displayed all the detected labels from the image. To try this technological marvel on your own phone, download the APK from Google Drive. The complete source code is in the GitHub repo. That’s it. Build this app, and a line of VCs will form right outside your door 😂.