Vision Demo

Simple application using Vision framework to detect the main object in an image.

Features

• capture a photo with a camera or choose one from library
• image is processed using Vision framework and pre-trained SqueezeNet model to detect the main object of the photo
• results with the most confidence percentage are shown in a form of tags
• tags can be selected or deselected to choose which ones to save
• image including tags is saved in Realm database
• all images can be seen on the gallery screen
• categories are shown on top of gallery screen and allow image filtering

Used libraries

• Vision for image processing
• Realm for data storage
• SnapKit for AutoLayout
• Combine for reactive programming