This is an android app for flowers recognition, which includes Android development, Convolution Neural Network and user database.
Although the AI accuracy achieved 88% accuracy, it is only applicable to the 102 flowers in the Oxford dataset. In other words, the capability of the app is unfortunately limited
because of my inability to web scrape flower images.
- Oxford 102 flowers dataset
- Modelled with MobileNetV2 architecture
- Stores user information with Firebase
- Logo drawn with Vectornator
Download from google play or this apk file in release.
under construction
Last updated: August 18, 2020
While using this App, it may ask the user to provide certain personally identifiable information that can be used to contact or identify the user. Personally identifiable information and corresponding access points may include, but is not limited to:
- Camera data -
- When the user choose to take photos from camera
- Files in photo gallery -
- When the user choose files from the photo gallery
- Username and email through Google sign in -
- When the user signs in or signs out this App using Google account
This app may use these data for the various purposes. The user's consent to this privacy policy means that the user agrees with the use of personal information for these purposes:
- Provide Services: to identify the flowers in the chosen photo.
- To manage User Account: to manage the user's registered account of the App. The Personal Data the user provides can give the user access to different functionalities of the App that are available to the user with a registered account.
The user's camera and photos data will not be retained in any way, however, the account information will be retained as long as this app exists on Github. This app and its developer will never share the user's information to any individual or entity, whereas the user has every right to transfer or delete stored personal data.
You entered a knowledge wasteland because no question has ever been asked ;(
Use Github Issues, send a message to aegerita or send an email to [email protected] to have your questions answered!