Use Google Cloud Vision API to analyze pictures uploaded in your application and guard against inappropriate content.
After obtaining your own Google API Key, create a Picguard instance to be used in the following examples:
let picguard = Picguard(APIKey: "foobar")If your app is a service which allows users to upload photos, you may often need a NSFW filter to reject pictures containing adult content, violence, spoof and medical content.
Picguard offers a unified method for checking all of above categories and uses an algorithm to calculate an average likelihood of NSFW content:
picguard.detectUnsafeContentLikelihood(image: anImage) { result in
switch result {
case .Success(let likelihood):
print("Likelihood of NSFW content: \(likelihood)")
case .Error(let error):
print("Error detecting NSFW content: \(error)")
}
}Sometimes you may need to know whether a photo contains any faces. For such case, Picguard offers a simple method to calculate a likelihood of any face being present in a picture:
picguard.detectFacePresenceLikelihood(image: anImage) { result in
switch result {
case .Success(let likelihood):
print("Likelihood of face presence: \(likelihood)")
case .Error(let error):
print("Error detecting face presence: \(error)")
}
}As Picguard is a fully featured Google Cloud Vision API client, you may compose your own requests that are not covered by above helpers and interpret the results your way:
picguard.annotate(image: anImage, features: [
.Face(maxResults: 2),
.Label(maxResults: 5),
.Landmark(maxResults: 3),
]) { result in
switch result {
case .Success(let response):
print("Face annotations: \(response.faceAnnotations)")
print("Label annotations: \(response.labelAnnotations)")
print("Landmark annotations: \(response.landmarkAnnotations)")
case .Error(let error):
print("Error analyzing image: \(error)")
}
}All *Annotation data types are richly documented and reflect equivalent API classes.
If you feel the need to play with Picguard before using it in your app, we prepared a demo playground especially for you! Just clone this repository, build the framework, open Picguard.playground and you're ready to go!
Picguard is written in Swift 2.2 and requires Xcode 7.3 or higher to be compiled. Minimum deployment target is iOS 9 and OS X 10.11.
If you're using Carthage, just add the following dependency to your Cartfile:
github "netguru/picguard-swift"
Using Picguard with CocoaPods is as easy as adding the following dependency to your Podfile:
use_frameworks!
pod 'Picguard'
Picguard is written in Swift 2.2 and requires Xcode 7.3 or higher to be developed.
Install Carthage dependencies using the following command and you're ready to go!
$ carthage bootstrap --platform 'iOS,Mac'Picguard follows standards described in GitHub's Swift Style Guide.
Please note that one of the points is to "use tabs, not spaces". Rationale: When using tabs, developers can configure the IDE to display them as 2, 3, 4 or any other number of spaces they like – thus there is no discussion on that topic.
Keep build settings in the appropriate xcconfig files inside Configuration directory. Please do not include any build settings in pbxproj.
Source files and spec files should be placed inside Sources and Tests directories respectively, without additional subfolders. You may use groups to manage the files inside the Xcode project itself.
Bitrise CI configuration should be described in bitrise.yml file and no workflows should be overwritten inside Bitrise.io UI.
This repository uses git-flow and protects develop and master branches from force pushes, and red builds, which means the whole development process is pull-request-driven.
This project is made with <3 by Netguru and maintained by:
Also, don't forget to check out the original Picguard, our ruby gem for validating uploaded images!
This project is licensed under the MIT License. See LICENSE.md for more info.