Visualize uses various neural networks and AI to identify photos and texts. The app is intended as visual aid for the visually impaired, while also offering a developer mode which allows the exploration of different AIs.
Interface Design:
Joseph Jin (General Interface Designer and Programmer)
Sam Hollenbeck (Graphics Design)
Vision neural networks:
-VGG16
Original Paper: Karen Simonyan & Andrew Zisserman. Keras Implementation: François Chollet
-Resnet50
Original Paper: Kaiming He and Xiangyu Zhang and Shaoqing Ren and Jian Sun. Keras Implementation: François Chollet
-Inceptionv3
Original Paper: Christian Szegedy, Vincent Vanhoucke, Sergey Ioffe, Jonathon Shlens, Zbigniew Wojna. Keras Implementation: François Chollet
-Using Microsoft Cognitive Services
http://go.microsoft.com/fwlink/?LinkID=829046.
Requests made using this service may be retained by Microsoft. For more information about the Cognitive Service Privacy, go to https://go.microsoft.com/fwlink/?LinkId=521839
WARNING:
This product should not be used in any situation where its failure could lead to the death or serious bodily injury of any person, or cause severe physical or environmental damage.