marți, 19 aprilie 2022

Facial Emotion Recognition

 

Human emotion detection is implemented in many areas requiring additional security or information about the person. Human emotions can be classified as: fear, contempt, disgust, anger, surprise, sad, happy, and neutral.

Facial Emotion Recognition is a technology used for analysing sentiments by different sources, such as pictures and videos.

FER typically has four steps. The first is to detect a face in an image and draw a rectangle around it and the next step is to detect landmarks in this face region. The third step is extracting spatial and temporal features from the facial components. The final step is to use a Feature Extraction (FE) classifier and produce the recognition results using the extracted features.

The FER 2013 dataset is usually used and also several  images for the five emotions are selected.(The emotions considered were happy, sad, angry, fear and neutral). These images are converted into NumPy arrays and landmark features are identified and extracted. A CNN model was developed with four phases where the first three phases had convolution, pooling, batch normalization and dropout layers. The final phase consists of flatten, dense and output layers

Convolutional neural network (CNN) is an algorithm of deep learning. Lecun first proposed its idea in 1989, and in 1998 proposed the application of this algorithm to handwritten digit recognition.  In 2012, Alex Krizhevsky won the Imagenet 2012 competition with CNN.

CNN can input image directly and get the final classification result without data preprocessing. By building a neural network model with a certain depth and combining nonlinear operations such as convolution and pooling, we can realize two important functions of imitating the hierarchical processing of human brain and local perception of visual nerve. It has been proved that the network has achieved good results in face recognition, speech recognition, vehicle detection and target tracking.  One role of a CNN is to reduce images into a form which is easier to process without losing features that are critical for good prediction.

According to some psychologists, communication occurring through facial expressions account for about 55% of communication, so machines can offer us more help if they are able to perceive and recognize human emotions.

 



Webliografie:

https://padlet.com/andreeaoprea99/w2otzumzrpqmhxnp

 

Niciun comentariu:

Trimiteți un comentariu

Neural network can read tree heights from satellite images

Neural network can read tree heights from satellite images Researchers at ETH Zurich have created a high-resolution global vegetation height...