In this paper, a new method is presented for estimating the gaze direction using Canonical Correlation Analysis (CCA), which finds a linear relationship between two datasets defining the face pose and the corresponding facial appearance changes. Afterwards, iris tracking is performed by blob detection using a 4-connected component labeling algorithm. Finally, a gaze vector is calculated based on gathered eye properties. Results obtained from datasets and real-time input confirm the robustness of this method.
Particle filters are often used for tracking objects within a scene. As the prediction model of a particle filter is often implemented using basic movement predictions such as random walk, constant velocity or acceleration, these models will usually be incorrect. Therefore, this paper proposes a new approach, based on a Canonical Correlation Analysis (CCA) tracking method which provides an object specific motion model. This model is used to construct a proposal distribution of the prediction model which predicts new states, increasing the robustness of the particle filter. Results confirm an increase in accuracy compared to state-of-the-art methods.
Emotion recognition currently is a very interesting subject in computer vision research, as we are heavily developing new methods of interaction between human and computer. This bachelor thesis is an introduction to emotion recognition using facial expressions and can be used as a guideline for implementing facial expression recognition. While an application was written and will be discussed, we also focus on other existing methods and explain them briefly.