A pre-trained model vs dedicated convolution neural networks for emotion recognition

Abstract

Facial expression recognition (FER) is one of the most important methods influencing human-machine interaction (HMI). In this paper, a comparison was made between two models, a model that was built from scratch and trained on FER dataset only, and a model previously trained on a data set containing various images, which is the VGG16 model, then the model was reset and trained using FER dataset. The FER+ data set was augmented to be used in training phases using the two proposed models. The models will be evaluated (extra validation) by using images from the internet in order to find the best model for identifying human emotions, where Dlib detector and OpenCV libraries are used for face detection. The results showed that the proposed emotion recognition convolutional neural networks (ERCNN) model dedicated to identifying human emotions significantly outperformed the pre-trained model in terms of accuracy, speed, and performance, which was 87.133% in the public test and 82.648% in the private test. While it was 71.685% in the public test and 67.338% in the private test using the proposed VGG16 pre-trained model.

Acoustic and visual geometry descriptor for multi-modal emotion recognition fromvideos

“…Consider two consecutive landmark points 𝑙 𝑖 , li and 𝑙 𝑖+1 . The Euclidean distance between them is calculated as (16). After that, they are normalized by performing division operations through the length as (17).…”

“…The recognition results are treated and exhibited on an information system as a website [15]. Using images from the internet will evaluate the models to discover the best model for recognizing human emotions and detecting faces [16]. The high-level motion feature frames are forwarded to the pre-trained CNN to distinguish the 17 emotions in the Geneva multimodal emotion portrayals dataset [17].…”

A Study on Emotion Analysis and Music Recommendation Using Transfer Learning

“…The model could detect emotions such as joy, sadness, surprise, fear, anger, contempt, and neutrality. Nawaf and Jasim (2023) compared two models: The VGG16 model, which was reset and trained using the FER dataset, and a model created from scratch and trained exclusively on the FER dataset. To determine which model would be most effective at identifying human emotions, the models were tested using photographs from the internet.…”