Sequential Feature Selection Using Hybridized Differential Evolution Algorithm and Haar Cascade for Object Detection Framework

Salefu Ngbede Odaudu, Emmanuel Adewale Adedokun, Ahmed Tijani Salaudeen, Francis Franklin Marshall,, Yusuf Ibrahim & Donald Etim Ikpe


Intelligent systems an aspect of artificial intelligence have been developed to improve satellite image interpretation with several foci on object-based machine learning methods but lack an optimal feature selection technique. Existing techniques applied to satellite images for feature selection and object detection have been reported to be ineffective in detecting objects. In this paper, differential Evolution (DE) algorithm has been introduced as a technique for selecting and mapping features to Haarcascade machine learning classifier for optimal detection of satellite image was acquired, pre-processed and features engineering was carried out and mapped using adopted DE algorithm. The selected feature was trained using Haarcascade machine learning algorithm. The result shows that the proposed technique has performance Accuracy of 86.2%, sensitivity 89.7%, and Specificity 82.2% respectively.

Full Text:



  • There are currently no refbacks.