Human Behavior Observation

Curious about the potential applications of automated, mass emotional analysis, Reallaer is committed to advancing the science and application of facial expression recognition. Reallaer maintains formal certification in Facial Action Coding System (FACS), developed by leading social science researchers in the facial expression analysis community.

Facial Expression
Analysis

Reallaer builds prototype technology to detect deception cues in personal interactions. These technologies enable viewers to understand intent, mood, interest, disinterest, and truthfulness through a person's facial expressions. Reallaer integrates multi-modal algorithms in a real-time system to signal detection of cues. Reallaer integrates automated feature extraction algorithms into real-time prototypes. We time synchronize multiple sensor data from academia and industry.

 

Case Study


Reallaer performs Independent Validation and Verification of systems and algorithms for facial recognition and video analytics. Reallaer focuses on characteristics associated with training data, and parameter tuning for video analysis. Our processes focus on evaluating algorithms for efficacy in real world operations. Algorithm optimization includes machine learning techniques with Hadoop/MApReduce parallel computing in a cloud environment to improve computing speed, model accuracy, and stability.

 

Automated Feature Extraction

Automated Feature Extraction

Remote
FACS Automation

Reallaer quickly and cost effectively evaluates the emotional responses of large numbers of individuals to specific triggers through remote FACS automation. Whether clients are developing television or internet advertising campaigns or crowd-sourcing concepts, Reallaer can provide emotional response research. Reallaer integrates behavioral theory with computer vision techniques to interpret facial expressions in real-time. Reallaer automated FACS classifier is built on a ever-growing data base of recognized and encoded facial expressions.

 

Case Study



Reallaer leads a multifunctional team to develop and instruct casual conversation training to improve data gathering in face-to-face exchanges.

 

 

Automated Intent Classification

Automated Intent Classification

Adaptive Learning to calibrate for baseline behaviors.