Umang Mehta - Emotion AI Developer Day 2016
Download as PPTX, PDF2 likes464 views
The document provides an overview of the Affectiva Emotion SDK for Windows, detailing its features, supported architectures, and software requirements. It includes instructions for getting started, examples of sample applications, and various use cases showcasing its application in different fields. Resources for further assistance, such as online documentation and community support links, are also included.
Umang Mehta - Emotion AI Developer Day 2016
- 1. Affectiva Emotion SDK for Windows Umang Mehta Software Engineer Affectiva
- 2. Agenda @affectiva • Overview • Getting Started • Architecture • SampleApplications • Live Demo • Use Cases • Helpful Resources • Questions
- 5. A range of emotions Valence: how positive or negative a person’s facial expressions Engagement: overall level and intensity of the emotion @affectiva
- 6. A person’s appearance Gender: identifies the human perception of gender expression Ethnicity: Person’s ethnicity @affectiva Age: Estimate age bucket (10 yrs) Glasses: presence of eye or sun glasses
- 7. And emojis Laughing Smiley Relaxed Wink Kissing Stuck Out Tongue Stuck Out Tongue and Winking Eye Scream Flushed Smirk Disappointed Rage @affectiva
- 8. Software and Hardware Requirements @affectiva Supported Processor Architecture • Windows 32 bit (x86) • Windows 64 bit (x64) Languages Supported • CSharp • C++ Hardware Requirement (Recommended) • Processor, 2 GHz • RAM, 1 GB • Disk Space (min) : 950 M Supported Operating System • Windows 7 and above
- 9. How to get started? @affectiva • Download the SDK - http://developer.affectiva.com/ • Run the Installer
- 11. Listeners/ CallBacks Image Listener • Face Listener Process Status Listener
- 12. Life-cycle of Detector Object @affectiva • Start the detector • Set the location of the model on disk • Set the classifiers • Process the file/ stream • Callback when something happens • Stop the detector
- 13. Sample Applications: @affectiva • AffdexMe - Analyzes input camera stream - Multi-face enabled - Small/Large Face Mode enable - Enabled Classifiers: Emotions, Expressions, Emojis, Gender and Glasses - Github : https://github.com/Affectiva/affdexme-win Download pre-build: http://developer.affectiva.com/sampleapps/ • csharp-sample-app - Analyzes a photo - Analyzes a recorded file - Analyzes input camera stream - Ability to switch Face Mode - Enabled Classifiers: Emotions, Expressions, Emojis, Gender, Glasses, Age and Ethnicity - Github: https://github.com/Affectiva/csharp-sample-apps
- 15. Live Coding Exercise Requirements: • IDE : Visual Studio 2015/ 2013 • AffdexSDK : http://developer.affectiva.com/ • Integrated Web Camera/ External Web Camera Camera Detector Demo
- 17. Cloverleaf Media, LLC Unique sensory concert experience Captures real time, crowd-sourced emotive information Displays data with a dazzling, interactive LED light show to transform the traditional concert experience Cloverleaf Media, LLC Gather insights shopper emotions, age, gender, ethnicity Personalized experience Ads in digital signage on retail shelves adapt to this data
- 19. @affectiva Helpful Resources • Online documentation: http://developer.affectiva.com • Q&A on StackOverflow: http://stackoverflow.com/questions/tagged/affdex-sdk • Chat on Gitter: https://gitter.im/Affectiva/sdk • Source code examples: https://github.com/affectiva?q=sample • Emotion-enabled apps: http://developer.affectiva.com/poweredbyaffdex/
- 21. Thank You!
Editor's Notes
- #5: Au02 -> Brow Raise Au26 -> Jaw Drop
- #10: The SDK is free for use for personal, open-source projects and for company whose revenue is less than $1M with few exclusions. Distributed using the Installer: affdex-native : C++ Namespace and header files Affdex : .Net Namespace Assembly Data: the classifier data folder required by both Affdex and affdex-native during runtime opencv_ffmpeg : A dynamic library which is required by Video Detector during runtime for video decoding
- #11: Camera – Analyzing a Camera Feed. We implement a Camera API that helps to get information about the camera, like camera started, isrunning, gives frames, stops the camera and so on Photo – Analyzing a Photo Video – Analyzing a recorded Video File Frame – Analyzing a Video Frame Stream is more client driven analyzing frame by frame. You can pass your own frame Detector : Base class which has all the methods like start, stop, process, setting the classifier
- #12: ImageListener – onImageResults() and onImageCapture() FaceListener – onFaceLost() and onFaceFound() ProcessStatusListener – onProcessingException() and onProcessingFinished
- #13: What needs to be done once you initialize a Detector object? Start detector Set the location of the data directory which comes with the SDK installer Set the classifiers you are interested in Process the camera stream/ the input files Give a Callback when an event happens Stop the Detector
- #15: Show the live demo.
- #16: How to integrate a
- #18: Experimental Marketing: Delivers double digit sales uplift via engaging interactive experience at point of decision Captures real time traffic, awareness and conversion Optimizes marketing messages by capturing measurable demographics and emotive data
- #19: Gaming: Real Time Interaction: Sense player emotion and adapt game in real time Insights & Analytics Play test player emotional engagement Gather emotion data for analysis in free to play game