What is it about?
The Object detection is method of computer vision for locate instances of objects in videos or images. Algorithms for object detection generally control Machine (ML) or Deep Learning (DL) for producing results. Several ML methods are used for detecting objects from image acquired through video. Huge incoming volume information is high for ML methods to manage. DL methods are used for recognizing objects and individuals from videos in accordance with handle high quantity of information as input. Several DL methods are used like Convolutional Neural Network (CNN), Recurrent Neural Network (RNN) and etc., to handle huge quantity of input data and also human and object recognition. Thus, Hybrid Whale optimization Algorithm (WOA) and Convolutional Neural Network (CNN) is proposed to detect object from video frames. WOA is used to enhance the method parameter of CNN architecture. Initially, the pre-processing technique is used to delete noise in image and quality of image is improved. Gaussian filter is utilized to subtract background in images. The proposed method improved accuracy of object detection and identifies the small targets in videos. The proposed algorithm attained high accuracy of 98.23%, precision of 96.78%, recall of 96.02% and f1-score of 96.45% which is higher than other existing methods.
Featured Image
Why is it important?
This research introduces a Hybrid Whale Optimization Algorithm (WOA) coupled with Convolutional Neural Network (CNN) for robust object detection in video frames, effectively enhancing accuracy and addressing small target identification. The proposed method achieves impressive performance metrics, surpassing existing techniques, with high accuracy (98.23%), precision (96.78%), recall (96.02%), and f1-score (96.45%).
Perspectives
Read the Original
This page is a summary of: Whale Optimization Algorithm with Convolutional Neural Network for Object Detection, November 2023, Institute of Electrical & Electronics Engineers (IEEE),
DOI: 10.1109/iciics59993.2023.10420963.
You can read the full text:
Contributors
The following have contributed to this page