Keypoint Density-based Region Proposal for Fine-Grained Object Detection and Classification using Regions with Convolutional Neural Network Features

TL;DR

This paper introduces a Keypoint Density-based Region Proposal method that significantly speeds up fine-grained object detection and classification using CNNs, making real-time applications more feasible without losing accuracy.

Contribution

The paper presents a novel region proposal technique, KDRP, that outperforms selective search in speed while maintaining accuracy for fine-grained object detection.

Findings

KDRP doubles the speed of detection and classification.

KDRP maintains comparable classification accuracy to existing methods.

Enables real-time fine-grained object detection using CNNs.

Abstract

Although recent advances in regional Convolutional Neural Networks (CNNs) enable them to outperform conventional techniques on standard object detection and classification tasks, their response time is still slow for real-time performance. To address this issue, we propose a method for region proposal as an alternative to selective search, which is used in current state-of-the art object detection algorithms. We evaluate our Keypoint Density-based Region Proposal (KDRP) approach and show that it speeds up detection and classification on fine-grained tasks by 100% versus the existing selective search region proposal technique without compromising classification accuracy. KDRP makes the application of CNNs to real-time detection and classification feasible.