Affdex Facial Coding

The challenge marketers face with consumer response...

Marketers recognize that emotion drives brand loyalty and purchase decisions. Yet, traditional ways of measuring emotional response - surveys and focus groups - create a gap by requiring viewers to think about and say how they feel. Neuroscience provides insight into how the mind works, but it typically requires expensive, bulky equipment and lab-type settings that limit and influence the experience.

   

Delivering true emotion insight

Affdex is an award-winning neuromarketing tool that reads emotional states such as liking and attention from facial expressions using an ordinary webcam...to give marketers faster, more accurate insight into consumer response to brands, advertising and media. It uses automated facial expression analysis recognition, also called facial coding, to analyze your face and interpret your emotional state. Offered as a cloud-based software-as-a-service, Affdex is a fast, easy and affordable to add into existing studies. MIT-spinoff Affectiva has some of the best and brightest emotion experts behind the Affdex platform science, providing the most accurate measurement today. This ongoing investment in research and development is focused not just on measuring, but also on predicting...which ads will really work to drive sales and build brands.

               See how Affdex works           Try Affdex Demo

 

Paul Ekman & Lie to MeHistory of Facial Coding. Foundation of Affdex

The research behind facial coding dates back to early studies by Darwin that concluded expressions are universal & are even shared with animals. Paul Ekman (1972) confirmed universality of 6 core expressions and popularized a facial action coding system (FACS) that is used to consistently describe facial expressions & movements. This system has received wide scale use, and has even been seen in the popular TV show, Lie to Me.

 

How Affdex Works

Cloud-Based Emotion Analytics:

By employing advanced computer vision and machine learning techniques that recognize and automate the analysis of tacit expressions, Affdex applies scientific methods to interpret viewers' emotional responses, quickly and at scale. Affdex delivers rich emotion insights –- Fast, Easy and Scalable, with just three simple steps.

__________________________________________________________________________________________________________________

 

Face Data Collection

Affdex is designed to use standard webcams found on everyday devices such as laptops, tablets, and mobile phones. Affdex can also be used with external webcams for in-lab and central venue settings. While the user watches online, streaming media, their individual expressions are captured via webcam. It’s unobtrusive, scalable and easy to collect emotion response.

Getting Started with Affdex Data Collection

Online or Offline: Affdex Anywhere is an offline data collection tool that is used for venues without reliable internet connection, or mobile.

Seamless Integration: Affdex is seamlessly embedded inside standard survey tools with a straightforward API (e.g. SPSS, ConfirmIt, SurveyGizmo) with a 1-time integration project; additional coding is not required for additional projects.

  • Affdex integration also includes a Camera Check page to improve data quality & optimize usable sessions, as well as an Opt-in to capture privacy consideration
 

__________________________________________________________________________________________________________________

     

Cloud Based Emotion Analysis

Once the face videos have been collected, they are streamed to the Affdex Cloud service for processing. In the cloud, Affdex reads emotional states such as surprise, dislike and attention from facial expressions. It employs advanced computer vision and machine learning techniques to recognize and automate the analysis of tacit expressions, and it applies scientific methods to interpret viewers' emotional responses quickly and at scale.

 

__________________________________________________________________________________________________________________

 

Affdex Analytics

The Affdex analytics provide overall emotion scores and real-time, scene-by-scene playback of facial data. See where viewers smile, which scenes capture the most attention and more, with our dynamic dashboard and benchmarked summary metrics.

The key deliverables in the Analytic Suite are:

In addition, Affdex projects are accessed through a self-service portal for quick & easy project setup, monitoring and analysis. The portal allows Affdex users to have:

  • Secure access to all current & past projects
  • An online project creation & configuration
  • Real-time data quality reports to monitor in field completion rates
 

__________________________________________________________________________________________________________________

 

Dynamic, Intuitive Dashboard

The dashboard provides interactive traces of key emotions experienced during the ad viewing and is fully integrated with your survey self-report data to ensure a cohesive and robust analysis for your clients. There are views for each emotion.

The emotion metrics include:

  • Net positivity (Valence)
  • Expressiveness
  • Attention
  • Concentration
  • Dislike
  • Smiles
  • Surprise

In addition to the dashboard metrics, Affdex includes several usability features that make creating client deliverables easier and more powerful. These include:

  • Face Finder: Exemplar faces are synchronized with the dashboard content to provide access to real, unfiltered panelist's reactions. This increases power & transparency of the results, adding qualitative detail to your quantitative study.
  • Bookmarks: Highlight and share key findings using bookmarks.
  • PowerPoint Friendly: Views make it easy and effective to share findings via Power Point.
 

__________________________________________________________________________________________________________________

 

Summary Metrics & Normative Benchmarks

Affdex Summary Metrics drive top-level analysis and
include comparison to regionally-specific norms.

 

Powering Emotional Insight


The core of the Affdex science is composed of patented emotion algorithms that are state-of-the-art, having been put to the test in over a thousand studies all over the globe. These machine learning, computer vision algorithms take face videos as inputs and provide frame-by-frame emotion metrics as outputs.

Hover over the image or click each section for details on the 3 pillars of our science

______________________________________________________________________________________________________________________________

Emotion Measures Beyond the Basic 6 Emotions

Affdex delivers both discrete and dimensional emotion metrics:

  • Discrete Metrics: Smile, Surprise, Dislike, Confusion
  • Continuous Metrics: Valence, Attention, Expressive

Affdex leverages our extensive facial video repository to guide and prioritize development of new measures, selecting those that are the most relevant in real life.

______________________________________________________________________________________________________________________________

Rigorous Science for Real World Use – Accuracy Global Validation

In order to produce highly accurate emotion metrics, Affdex classifiers are trained on both posed and spontaneous, naturally occurring data. In addition to training, we also test our classifiers on both posed and spontaneous data. We insist that our classifiers perform at 80-90% or better to reach production use.

 

Validation studies have also been completed in Asia, Latin America, US and UK to confirm that key facial expressions are universal. In these studies, we also uncovered that while expressions were universal, the magnitude of the expressions varied greatly. This emphasized the need for market-specific normative data, which is part of every Affdex study.

 

______________________________________________________________________________________________________________________________

Continuous Improvement - Platform for classifier training & validation

face database

Framework for labeling & training new AU's to improve accuracy (repeatable & scalable classifier creation & testing)

 

Largest repository of spontaneous facial expressions (over 50,000,000 real world face frames)

 

Significant investment in science infrastructure, a dedicated labeling team and automation for classifier creation and testing

News

  • News Article | Posted on 12/23/2013

    2013 Review: The biggest success stories

    The reviews of 2013 so far have had some common themes, with big data and mobile featuring heavily. These topics also cropped up in contributors’ views on the biggest success stories of the year, but there was plenty of variation besides. Another technology-related shout-out went to Affectiva. “Affectiva’s facial coding technology has got to be one of the biggest winners this year. Adopted by Millward Brown as a standard feature in their global link tests and presenting at every conference going.” Jon Puleston, GMI

  • News Article | Posted on 12/10/2013

    Facial Coding Technology Helps Marketers Optimize Ad Effectiveness

    Facial expressions can reveal what people do not say. And that is the premise that the Millward Brown Link with Affdex Facial Coding technology uses to gauge viewers’ true reactions to pre-launched advertisements.

  • News Article | Posted on 10/28/2013

    Startup Gets Computers to Read Faces, Seeks Purpose Beyond Ads

    A technology for reading emotions on faces can help companies sell candy bars. Now its creators want to see whether it can take on bigger problems in areas including education.

  • News Article | Posted on 5/24/2013

    One Day Your Phone Will Know If You’re Happy or Sad

    But what if these devices could really read our emotions? What if they could interpret every little gesture, every facial cue so that they can gauge our feelings as well as–maybe better than–our best friends? And then they respond, not with information, but what might pass for empathy.

  • News Article | Posted on 4/8/2013

    Affectiva presenting at Esomar Congress 2013!

    "Ads that evoke emotions are more entertaining and memorable, but do they really drive product sales? Using Facial Coding to Understand The Relation between Emotional Ads and Sales Effectiveness.

    In a recent study, Affectiva's world’s largest data repository ties emotions in ads to real-world sales effectiveness. Using facial coding, their researchers crowdsource emotional responses of 1000+ viewers to over 140 ads in 4 countries and across 4 product categories - distilling sales data associated for each ad. They build and validate models that identify the facial measures and emotion trajectories that are most predictive of sales performance.

  • News Article | Posted on 2/20/2013

    For the first time, brands will be able to understand and react in real-time to users’ emotional responses to their online advertising

    TheMarketingblogAffdex uses Computer Vision and Machine Learning algorithms, developed in partnership with MIT, to detect facial expressions and head gestures obtained from webcams or mobile cameras. It assesses, analyses and interprets the user’s reactions to content to detect the full range of emotions from joy, discomfort and indifference to rapt engagement.

  • News Article | Posted on 2/19/2013

    Affectiva Inks Deal With Ebuzzing Social To Integrate Face Tracking & Emotional Response Into Online Video Ad Analytics

    The deal will mean that companies like Heineken and Red Bull will be able to track in real time how users are responding to their advertisements in the wild. That will not only mean taking away the need to run special market research sessions on limited groups of users, but potentially introduces new ways of measuring how effective a video has been, moving away from more traditional online ad metrics like page views and dwell time.

  • News Article | Posted on 2/2/2013

    Building better Super Bowl ads by watching you watch them

    A 3-year-old company takes technology from MIT's Media Lab and applies it to ad testing. Welcome to the future of advertising, where the wisdom of spending a reported $4 million for a 30-second spot in the Super Bowl doesn't have to be left to the imagination of an ad agency's creative team and the honesty of focus groups.

  • News Article | Posted on 1/29/2013

    Creating the Perfect Super Bowl Ad

    Are advertisers striking the right balance between entertaining and promoting their brands? Could companies entertain less and get consumers to buy?

  • News Article | Posted on 1/18/2013

    Let the NeuroGames Begin

    Neurogaming is set to disrupt the gaming industry and transform the gaming experience over the next 5 yrs. Companies, like Affectiva, are using a variety of techniques such as analyzing facial expressions, to understand emotional reactions within the game design process.