Pet dog facial expression recognition based on convolutional neural network and improved whale optimization algorithm

Abstract

Pet dogs are our good friends. Realizing the dog’s emotions through the dog's facial expressions is beneficial to the harmonious coexistence between human beings and pet dogs. This paper describes a study on dog facial expression recognition using convolutional neural network (CNN), which is a representative algorithm model of deep learning. Parameter settings have a profound impact on the performance of a CNN model, improper parameter setting will make the model exposes several shortcomings, such as slow learning speed, easy to fall into local optimal solution, etc. In response to these shortcomings and improve the accuracy of recognition, a novel CNN model based on the improved whale optimization algorithm (IWOA) called IWOA–CNN is applied to complete this recognition task. Unlike human face recognition, a dedicated face detector in Dlib toolkit is utilized to recognize the facial region, and the captured facial images are augmented to build an expression dataset. The random dropout layer and L2 regularization are introduced into the network to reduce the number of transmission parameters of network and avoid over fitting. The IWOA optimizes the keep probability of the dropout layer, the parameter λ of L2 regularization and the dynamic learning rate of gradient descent optimizer. Carry out a comparative experiment of IWOA–CNN, Support Vector Machine, LeNet-5 and other classifiers for facial expression recognition, its results demonstrate that the IWOA–CNN has better recognition effect in facial expression recognition and also explain the efficiency of the swarm intelligence algorithm in dealing with model parameter optimization.

Human Action Recognition System Using LRCN

“…Experimental evaluations are conducted using benchmark datasets like Pascal VOC, MPII, LSP, and a self-collected dataset (SCAID), showcasing the SSD-ResNet50-Person model's improved accuracy in human detection and the CPM-Stage 6 model's proficiency in skele-ton keypoints detection. The research demonstrates signifi-cant advancements in action recognition accuracy, particularly through the [1] CNN-8 network's performance on different datasets, while also identifying challenges such as the lower detection accuracy of extremity joints and classification errors between similar actions.…”

“…©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | [1]The paper underscores the importance of HAR in various domains including healthcare, surveillance, and human-computer interaction, and discusses recent advancements in HAR research, including the adoption of deep learning models, hybrid approaches, and the use of wearable sensors to improve accuracy and model performance. Additionally, it delves into the technical workings of CNNs, explaining the functions of convolutional and pooling layers in feature extraction and the role of fully connected layers in learning high-level features for accurate activity classification.…”

“…To advance this field, innovative approaches have emerged that integrate cutting-edge technologies and advanced machine learning techniques. [1] One such approach is the development of a sophisticated Human Action Recognition System by combining Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks. By leveraging CNNs to capture spatial features and LSTMs to model temporal dependencies, this system aims to create a robust model capable of accurately recognizing a diverse array of human actions with high precision.…”

Double strategy fusion improved sparrow search algorithm

“…The introduction of Brownian motion into the previous position update formula of SSA algorithm can effectively improve the global exploration ability of the algorithm. The standard Brownian motion formula is as follows (7):…”

“…At present, particle swarm algorithm (PSO) 5 , grey wolf optimization (GWO) 6 , whale optimization algorithm (WOA) 7 and so on are widely used and common. These algorithms provide more ideas for solving complex optimization problems.…”