Mobile image search system

A method comprising:

• capturing, by an image sensor of a mobile device, image data for a first image;
• processing, by the one or more processors, the image data using a filter convolved across a width and a height of the first image in a neural network to determine one or more characteristics of the image data selected from a set of tags and based at least in part on image content of the first image, the neural network comprising a first layer and a second layer in a first subgraph;
• determining that the first and second layers in the first subgraph comprises an output of the neural network;
• associating each tag of the set of tags with the output;
• storing the one or more characteristics of the image data in the content database;
• processing a user input requesting a user interface output on the display;
• presenting the user interface with a first plurality of images comprising the first image, wherein the plurality of images are based on a set of default search content characteristics from the content database;
• and
• responsive to determining that the first and second layer do not comprise the output, reserving a location in the memory associated with the image data comprising the one or more characteristics.

A non-transitory computer readable medium comprising instructions that, when executed by one or more processors of a mobile device, cause the mobile device to perform operations comprising:

• accessing the image data for a first image;
• processing, by the one or more processors, the image data using a filter convolved across a width and a height of the first image in a neural network to determine one or more characteristics of the image data selected from a set of tags and based at least in part on image content of the first image, the neural network comprising a first layer and a second layer in a first subgraph;
• determining that the first and second layers in the first subgraph comprises an output of the neural network;
• associating each tag of the set of tags with the output;
• storing the one or more characteristics of the image data in the content database;
• processing a user input requesting a user interface output on the display;
• presenting the user interface with a first plurality of images comprising the first image, wherein the plurality of images are based on a set of default search content characteristics from the content database;
• and
• responsive to determining that the first and second layer do not comprise the output, reserving a location in the memory associated with the image data comprising the one or more characteristics.