An Improvement of Object Detection Performance using Multi-step Machine Learnings

TL;DR

This paper proposes a multi-step machine learning pipeline for object detection, introducing a calibration model that leverages contextual information to improve accuracy.

Contribution

It introduces a novel calibration model within a multi-step pipeline, enhancing object detection performance by integrating domain knowledge and contextual cues.

Findings

Improved average precision by 0.8-1.9 points over existing detectors

Utilized a convolutional neural network for calibration

Enhanced explainability and accuracy of object detection

Abstract

Connecting multiple machine learning models into a pipeline is effective for handling complex problems. By breaking down the problem into steps, each tackled by a specific component model of the pipeline, the overall solution can be made accurate and explainable. This paper describes an enhancement of object detection based on this multi-step concept, where a post-processing step called the calibration model is introduced. The calibration model consists of a convolutional neural network, and utilizes rich contextual information based on the domain knowledge of the input. Improvements of object detection performance by 0.8-1.9 in average precision metric over existing object detectors have been observed using the new model.