eversvd,+5 (17)

Object detection is a prominent application within the field of computer vision, where the primary aim has shifted not only to enhance accuracy but also to develop lightweight models. This paper introduces a novel lightweight object detection system that bypasses the need for object classification, employing clustering methods instead. Specifically, we propose a methodology that utilizes the k-means clustering algorithm for effective object detection. The number of objects is determined using the elbow method, making the approach both efficient and accessible for applications that do not necessitate object classification. Our methodology is thoroughly explained and supported by graphical representations, providing clarity on its functioning. The proposed system can demonstrate significant advantages, particularly in scenarios that prioritize speed and resource conservation over the complex requirements of traditional deep learning-based object detection techniques. Through this work, we aim to contribute to the ongoing discourse on balancing performance and efficiency in object detection tasks, showcasing how simpler methodologies can still yield valuable results in practical applications.