The Facial Emotion Recognition (FER-2013) Dataset for Prediction System of Micro-Expressions Face Using the Convolutional Neural Network (CNN) Algorithm based Raspberry Pi

Zahara, Lutfiah; Musa, Purnawarman; Wibowo, Eri Prasetyo; Karim, Irwan; Musa, Saiful Bahri

doi:10.1109/icic50835.2020.9288560

Cited by 101 publications

(39 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An accuracy of 71% was achieved for neutral facial expression, whilst lower accuracies were achieved for angry and scared facial expression; the model tended to mistake them for each other. However, this result is consistent with recent findings in the field [ 41 , 42 ]. The overall accuracy of our model on the FER2013 dataset was 69%, which is considered higher than the human-level accuracy of 65%.…”

Section: Discussionsupporting

confidence: 94%

Improving Human–Robot Interaction by Enhancing NAO Robot Awareness of Human Facial Expression

Filippini

Perpetuini

Cardone

2021

Sensors

View full text Add to dashboard Cite

An intriguing challenge in the human–robot interaction field is the prospect of endowing robots with emotional intelligence to make the interaction more genuine, intuitive, and natural. A crucial aspect in achieving this goal is the robot’s capability to infer and interpret human emotions. Thanks to its design and open programming platform, the NAO humanoid robot is one of the most widely used agents for human interaction. As with person-to-person communication, facial expressions are the privileged channel for recognizing the interlocutor’s emotional expressions. Although NAO is equipped with a facial expression recognition module, specific use cases may require additional features and affective computing capabilities that are not currently available. This study proposes a highly accurate convolutional-neural-network-based facial expression recognition model that is able to further enhance the NAO robot’ awareness of human facial expressions and provide the robot with an interlocutor’s arousal level detection capability. Indeed, the model tested during human–robot interactions was 91% and 90% accurate in recognizing happy and sad facial expressions, respectively; 75% accurate in recognizing surprised and scared expressions; and less accurate in recognizing neutral and angry expressions. Finally, the model was successfully integrated into the NAO SDK, thus allowing for high-performing facial expression classification with an inference time of 0.34 ± 0.04 s.

show abstract

Section: Discussionsupporting

confidence: 94%

Improving Human–Robot Interaction by Enhancing NAO Robot Awareness of Human Facial Expression

Filippini

Perpetuini

Cardone

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…The classification dataset is taken from Kaggle, an online provider of datasets. Expression data is divided into 7 basic expressions (Meng, Liu, Cai, Han, & Tong, 2017), (Zahara, Musa, Prasetyo Wibowo, Karim, & Bahri Musa, 2020). Datasets are used for the training and validation process on the machine learning model in classifying emotions.…”

Section: Dataset Collectionmentioning

confidence: 99%

Alexnet Architecture and Fuzzy Analysis on Talent Judge Decision Prediction Based on Facial Expression

2022

View full text Add to dashboard Cite

The expression on the human face is a means of non-verbal communication. In the talent search event, the facial expressions shown by the judges when watching the participants' performances became one of the components to see whether the contestant who was performing could qualify for the next round or he would fail. Haar cascade is used to provide the location of the face in the frame and to classify the expressions on the face, a CNN model with modified AlexNet architecture is used which increases the accuracy by 5% from the original alexnet. A fuzzy Algorithm is used to predict the judge's decision based on how many facial expressions appear during the participant's appearance. The decision prediction system for talent search judges based on facial expressions using fuzzy is considered effective in predicting decisions, after being tested the system can predict decisions with an accuracy rate of 83%.

show abstract

“…Returning to CNNs, the basic idea is that the many levels of the neural network learn properties that are specific to a particular emotion. Because the filters don't reveal a lot of information [7].…”

Section: Fully Connected Layermentioning

confidence: 99%

“…Face emotion detection software is used to make it easier to recognize and verify individual emotions based on their facial features as a result, understanding facial features and their behavior is extremely important [6]. Facial expression is the outcome of a facial gesture or expression that demonstrates the position of the muscles on the human face as a type of nonverbal communication as well as an important means of conveying one's emotions to others in the form of sentiments, intentions, and views [7]. The main task in facial expression is to map various facial expressions to their corresponding emotional states.…”

Section: Introductionmentioning

confidence: 99%

Face Recognition Emotions Detection Using Haar Cascade Classifier and Convolutional Neural Network

Rasheed

Khadam

Majeed

et al. 2022

Preprint

View full text Add to dashboard Cite

Nowadays, Face recognition and emotion detection are the most pop- ular area of research. In the past two decades, many applications related to facial emotions have been developed and different methods are used to detect human facial emotions. The goal of this research is to detect human facial emotions i.e. happy, sad, disgust, anger, fear, neutral, or surprise. In this paper, we proposed Haar Cascade classi- fier for face detection emotions detection. The Haar-like features have different types, the edge feature is used to detect edges around the object while the line and rectangle features are used to detect the slanted line of the object. The proposed technique recognizes facial emotions through deep-learning- convolutional neural networks based on FER2013. The FER2013 dataset contains 35887 entries of seven facial emotions and is categorized as 28709 training images, 3589 vali- dation images, and 3589 test images. The proposed method consists of six convolutional layers, three max-pooling layers, and four fully con- nected layers. The output layer is softmax is used to calculate the result from the convolutional neural networks. Through the experimental results, our proposed method achieved a validation accuracy of 65.59 %.

show abstract

The Facial Emotion Recognition (FER-2013) Dataset for Prediction System of Micro-Expressions Face Using the Convolutional Neural Network (CNN) Algorithm based Raspberry Pi

Cited by 101 publications

References 9 publications

Improving Human–Robot Interaction by Enhancing NAO Robot Awareness of Human Facial Expression

Improving Human–Robot Interaction by Enhancing NAO Robot Awareness of Human Facial Expression

Alexnet Architecture and Fuzzy Analysis on Talent Judge Decision Prediction Based on Facial Expression

Face Recognition Emotions Detection Using Haar Cascade Classifier and Convolutional Neural Network

Contact Info

Product

Resources

About