Enhancing real-time object detection and recognizing immediate changes in object size - a review of advances in computer vision model

Over the past decade, there has been a significant surge in the utilization of real-time object detection applications. In a wide range of industries, including robots, augmented reality, driverless cars, medical field, and video surveillance, real-time object detection has become essential. The various computer vision combined with deep learning framework has distinguished itself among the many object identification algorithms with its exceptional balance of speed and accuracy. Facilitating the accurate and quick identification of items in pictures. In this paper we provide critical review with analysis and evolution of the popular computer vision model framework, explaining the main advancements, modifications, and enhancements made in each iteration, starting with YOLO and ending with Cascade R-CNN. After outlining all present computer vision models are limited to object classification, detection, segmentation with pose estimation only. we summarize important aspects and explore potential prospects with suggested improved deep learning framework with common metrics and preprocessing using an improved computer vision framework along with an open-source library. This integration enables not only detecting objects, but also tracking changes in real-time and making accurate measurements to meet the industrial demand for computer vision applications.