A Picture and A Story

Hi guys,
I welcome you all to my blog today. It's a beautiful day of the week, and I am delighted to share with you my activities for the day, which the picture below clearly explains. Today, being 27-08-2025, is the day set aside for our presentation for the last module we are working on before we proceed to our dissertation.

I worked on a very interesting topic, which I defended during the presentation. My topic is Efficient Hybrid CNN-GNN with Attention-Based Feature Selection for Facial Expression Recognition. Here, we tend to use two different models, which are CNN and GNN, for our analysis alongside attention mechanisms for explainability.

Let me discuss with you all briefly about the result obtained in the analysis or what the topic is generally all about. The research is all about computer vision, where a computer can look at the face of a human and then determine the emotion of the person based on facial expression.

Based on my analysis, I noticed that the system easily identifies some expressions but is confused when other expressions, which are not common, are used. Based on the findings, I concluded that a lot needs to be added when dealing with FER, such as more data on the expressions that the system finds difficult to predict, and also improve the attention mechanism so that it becomes easier and better to recognize and predict.

So that is all about the picture you have seen above. I'm going to stop here, and we are going to see tomorrow as I share with you all yet another beautiful picture.