FACIAL EMOTION DETECTION

 

Hello reader, I want to share my experience of doing a capstone project at Bennett University through this article, I did this project with my teammate Abhinav under the guidance of Dr. Suneet Gupta and Dr. Deepak Garg.

Emotion Detection using Facial Expression, I decided to do this project as it is an integration of all the technical stuff I learned in my college days. Our project covers many topics like OpenCV for Image Processing and Video processing, Deep Convolutional Neural Networks for recognizing facial points. We are using the website as an intermediate between the user and our trained model, so to achieve it we need to establish a Bidirectional connection between the user and the client, so we are implementing this connection by using flask( a python framework). Coming to the library section we are using TensorFlow, Keras, sklearn, pandas, etc.

In our project. We build a CNN model which analyses the facial points and it estimates the emotion out of those facial points, and that CNN model gives the emotion of the face. And this CNN model was integrated with the web application, where the web application collects the face from the user and as it is integrated with the CNN model, the CNN model gives the emotion and that emotion is displayed on the web application. We took four months to complete the project.

Challenges we faced:

1. In the process of detection light matters to identify the facial points.

2. While training the deep learning model emotion detection we have to make sure that the data set does not contain blurred images. while testing the poor quality of the camera is not able to detect the emotion accurately.

3. In the case of covering the facial points it is difficult to detect the emotion of the person.

4. while training I took a long time to complete.

5. while integrating the CNN and web application I faced many issues which are some times web application is not able to communicate with the CNN model.

Resources used in our project:

1. For the process of training the deep learning models for emotion detection we used Data Sets.

2. For Our Project we used Open CV for analyzing the image like facial detection.

3. We used the CNN method which is used to predict the facial expression of the input image.

4. For the backend we used the Numpy package which helps to convert the images into arrays.

5. To integrate the CNN model and web application I used a flask.

Finally, it is a wonderful experience in BTECH life. I would like to thank Bennett University, Dr. Suneet Gupta, and Dr. Deepak Garg for their support.