Object detection based on template matching through use of best-so-far ABC

Area of Science:

• Computer Vision
• Artificial Intelligence
• Optimization Algorithms

Background:

• Swarm intelligence (SI) algorithms, like the artificial bee colony (ABC), are increasingly used for complex optimization tasks.
• Existing ABC algorithms show potential but can be improved for efficiency and solution accuracy.
• Object detection remains a computationally intensive task, necessitating optimized algorithms.

Purpose of the Study:

• To apply the advanced best-so-far artificial bee colony (ABC) algorithm to object detection.
• To utilize RGB level histogram differences between target and template objects as the objective function for optimization.
• To evaluate the performance of the best-so-far ABC approach in terms of detection accuracy and computational time.

Main Methods:

• Implementation of the best-so-far ABC algorithm, a modified version of the standard ABC.
• Object detection framework based on template matching.
• Objective function defined by the difference between RGB histograms of the target and template objects.

Main Results:

• The best-so-far ABC algorithm demonstrated superior performance compared to ordinary ABC and other state-of-the-art ABC variants.
• The proposed method successfully achieved accurate object detection.
• Significant optimization in the time required to reach detection solutions was observed.

Conclusions:

• The best-so-far ABC-based approach offers a promising method for efficient and accurate object detection.
• RGB histogram comparison combined with swarm intelligence optimization provides an effective objective function.
• This research contributes to advancing SI algorithms for practical computer vision applications.