Fast Learning and Prediction for Object Detection using Whitened CNN Features

TL;DR

This paper presents a method that combines pre-trained CNN features with a fast linear classifier to achieve high-performance, real-time object detection with minimal training data, suitable for efficient sliding-window detection.

Contribution

It introduces a novel approach integrating CNN features with Exemplar-LDA for rapid learning and detection, enhancing real-time object detection capabilities.

Findings

High detection accuracy with few training samples

Fast model training and inference

Effective sliding-window detection in real-time

Abstract

We combine features extracted from pre-trained convolutional neural networks (CNNs) with the fast, linear Exemplar-LDA classifier to get the advantages of both: the high detection performance of CNNs, automatic feature engineering, fast model learning from few training samples and efficient sliding-window detection. The Adaptive Real-Time Object Detection System (ARTOS) has been refactored broadly to be used in combination with Caffe for the experimental studies reported in this work.