In this project, the goal is to apply a Deep Learning model based on Convolutional Neural Networks (CNN) to recognize emotions (anger, disgust, fear, happiness, neutrality, sadness); gender (female, male), and age group (middle age, old, young) to which an image belongs to. The team will propose the architecture and find the hyperparameters of the model. Finally, a solid experimentation must be conducted to achieve an accuracy greater than 70%.
The dataset contains images of individuals with varying emotions, gender, and age groups. It is composed of a total of n images, with m images for each category (emotion, gender, and age group). The images are of size width x height x channels where channels refers to the number of color channels (usually 3 for RGB). The dataset will be split into training, validation, and testing sets.