This Small Business Innovation Research (SBIR) Phase I project aims to develop Affectiva Face Reader Platform; a cloud-enabled software as a service platform for the analysis of emotional and cognitive states from the face. Affectiva Face Reader Platform addresses potentially lucrative business opportunity in: 1) market research, 2) media research, 3) product testing and 4) usability testing, offering insights into customer resonance. The approach consists of building a multitier architecture that makes facial expression analysis seamless, scalable and affordable.
The company envisions a product that offers face-analysis as an affordable software as a service solution using highly scalable cloud-computing resources, enabling use by academics and smaller users, and higher-profit use by heavy users in industry. The goal is a technology service that positively impacts the way customers and businesses communicate about product experiences. The proposed product is potentially transformative in several ways: the company is allowing more accurate understanding of an important aspect of human communication, and they are democratizing market research. In addition, the product if successfully deployed has the potential to accelerate psychological and clinical research on social intelligence.
Every year, billions of dollars are spent trying to understand how consumers truly feel and how these feelings drive product perception and purchasing. Emotions are currently captured on questionnaires or in focus groups, both of which are inaccurate and do not predict outcomes in the market. What is needed is the ability to capture the emotions of customers naturally, while people engage with content, products and services, without interrupting their experience, and to do so in a way that improves business’s ability to make important product-related decisions. The innovation we proposed for this SBIR Phase I is the development of the Affectiva Facial Analysis Platform (Affdex), the first-in-the-world cloud-enabled software-as-a-service platform for the automated analysis of facial expressions and emotional states. Using input from a webcam and only with the person's permission, Affdex uses computer vision techniques to track the facial expressions, such as smiles, eyebrow flashes and attention. Affdex disrupts the current media measurement market by capturing new, precise, behavioral metrics around how viewers feel about content in real-time and at scale. In less than 6 months, Affdex has already crowd-sourced more than 9 million facial images from 12,000 sessions of people watching content and now hosts the largest database of people’s emotional response to media, allowing precise ranking and normative benchmarking. Intellectual merit: Our team combines strong scientific expertise with seasoned business experience. Over the past 10 years, our team at MIT has led research in automatic facial expression analysis. The technology development was originally funded by NSF to develop an intervention that helps individuals on the autism spectrum with recognition of facial emotions. These funds brought the PI of this SBIR (Rana el Kaliouby) to MIT as an international post-doc in Rosalind Picard’s Affective Computing Group at MIT, which later led to them co-founding Affectiva. David Berman is Affectiva’s chief executive officer. Prior to Affectiva, David served as president of worldwide sales and services at WebEx Communications, a Cisco company, where he implemented the software-as-a-service sales and marketing model. which was critical to the company’s successful initial public offering in 2000 and noted acquisition by Cisco for $3.2 billion in 2007. Broader impacts: Our long-term goal is to bridge the communication gap between companies and their customers. This is potentially transformative in many ways: not only are we allowing more accurate understanding of an important aspect of human communication, we are also democratizing market research. Typically market research has only been available to Fortune 100 companies who can afford to spend millions on market research. We aim to offer an affordable software as a service solution using highly scalable cloud-computing resources. A small startup can afford to pay per emotion, or pay per usage, without having to hire an expensive market research team. Additionally, because market researchers waste resources producing products that people politely say they want (even when they don’t), our work has the potential to also save many valuable natural and human resources by helping businesses more accurately identify what customers want. Affdex is also potentially transformative for science - enabling many more researchers to measure facial expressions of emotion, helpig them answer key questions about how humans communicate with each other.
