RCNN_ensemble.md

RCNN -> SPPnet -> Fast-RCNN -> Faster-RCNN

R-CNN

R-CNN Rich feature hierarchies for accurate object detection and semantic segmentation

CVPR 2014 code:http://www.cs.berkeley.edu/~rbg/rcnn

Abstract

• apply high-capacity CNNs to bottom-up region proposals in order to localize and segment objects.

• when labeled training data is scarce, supervised pre-training for an auxiliary task, followed by domain-specific fine-tuning, yields a significant performance boost.

Introduction

Object detection with R-CNN

Module 1: generates category-independent region proposals
Module 2: a large convolutional neural network that extracts a fixed-length feature vector from each region
Module 3: a set of classspecific linear SVMs

SPPnet

Fast R-CNN

Faster R-CNN