Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

A UIImageView clone that adjusts image content to show faces.

branch: master
readme.md

FaceImageView

A UIImageView clone with a catch:

FaceImageView automatically scales its image to fill the bounds and keep any faces detected in view. This can be useful if you want to display people-based dynamic content with UIViewContentModeScaleToFill but are having issues with it displaying the wrong part of the image, like the background, or someone's feet.

The class uses the CoreImage face detection APIs available in iOS 5.0 or later. This API works in many clear cases, but is not all that reliable. If no faces are detected in an image, FaceImageView defaults to centering the image, like UIImageView's behavior with UIViewContentModeScaleToFill.

A small demo app is included to try out different images and view sizes.

Usage

Installing

Add FaceImageView.{h,m} included in the demo project to your source, or, using CocoaPods, add the following to your Podfile:

pod 'FaceImageView'

Using

As simple as instantiating a FaceImageView object, giving it an image, and adding it to your view:

FaceImageView *imageView = [[FaceImageView alloc] initWithFrame:<CGRect>];
imageView.image = <UIImage>;
[self.view addSubview:imageView];

It may also be wise to set the image asynchronously, as face detection can sometimes take a second:

[imageView setImage:<UIImage> completion:^{}];

Demo App

Watch faces automatically stay centered in the view as it changes size:

demo


Further thoughts:

  • Have it give suggestions for minimum height, as to not entirely cut off a face?
  • With multiple faces, maybe prefer to keep at least one face entirely in view rather than halfway between many?
  • It seems like face detection implementation has changed between iOS 6 and 7. The former detects the face in image #4, but the latter does not. Not sure what to do with this, maybe it is temporary.

Posted for the Objective-C hackathon effort :)

Something went wrong with that request. Please try again.