Object Detection With Deep Learning: A Review (2018-07-15T00:00:00.000000Z)

TL;DR

This paper provides a review of deep learning-based object detection frameworks and focuses on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further.

Abstract

Due to object detection’s close relationship with video analysis and image understanding, it has attracted much research attention in recent years. Traditional object detection methods are built on handcrafted features and shallow trainable architectures. Their performance easily stagnates by constructing complex ensembles that combine multiple low-level image features with high-level context from object detectors and scene classifiers. With the rapid development in deep learning, more powerful tools, which are able to learn semantic, high-level, deeper features, are introduced to address the problems existing in traditional architectures. These models behave differently in network architecture, training strategy, and optimization function. In this paper, we provide a review of deep learning-based object detection frameworks. Our review begins with a brief introduction on the history of deep learning and its representative tool, namely, the convolutional neural network. Then, we focus on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further. As distinct specific detection tasks exhibit different characteristics, we also briefly survey several specific tasks, including salient object detection, face detection, and pedestrian detection. Experimental analyses are also provided to compare various methods and draw some meaningful conclusions. Finally, several promising directions and tasks are provided to serve as guidelines for future work in both object detection and relevant neural network-based learning systems.

Authors

Zhong-Qiu Zhao

1 papers

Shou-tao Xu

1 papers

Xindong Wu

2 papers

TL;DR

Abstract

Authors

References252 items

Training

Face recognition from multiple stylistic sketches: Scenarios, datasets, and evaluation

Low-Complexity Approximate Convolutional Neural Networks

Multiple Pedestrian Tracking From Monocular Videos in an Interacting Multiple Model Framework

Saliency-based Sequential Image Attention with Multiset Prediction

Stepwise Metric Promotion for Unsupervised Video Person Re-identification

Object Detection Meets Knowledge Graphs

Pedestrian Detection Based on Fast R-CNN and Batch Normalization

DSOD: Learning Deeply Supervised Object Detectors from Scratch

Jointly Attentive Spatial-Temporal Pooling Networks for Video-Based Person Re-identification

Non-local Deep Features for Salient Object Detection

Mimicking Very Efficient Network for Object Detection

RON: Reverse Connection with Objectness Prior Networks for Object Detection

Embedded Streaming Deep Neural Networks Accelerator With Applications

Scale-Aware Face Detection

Face Detection through Scale-Friendly Deep Convolutional Networks

Face R-CNN

Real Time Image Saliency for Black Box Classifiers

Learning Detection with Diverse Proposals

Pixelwise Instance Segmentation with a Dynamically Instantiated Network

Unsupervised Learning from Video to Detect Foreground Objects in Single Images

Mask R-CNN

Deformable Convolutional Networks

Object Detection in Videos with Tubelet Proposal Networks

Face Detection using Deep Learning: An Improved Faster RCNN Approach

DSSD : Deconvolutional Single Shot Detector

YOLO9000: Better, Faster, Stronger

Finding Tiny Faces

Feature Pyramid Networks for Object Detection

To boost or not to boost? On the limits of boosted trees for object detection

Speed/Accuracy Trade-Offs for Modern Convolutional Object Detectors

Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields

Fully Convolutional Instance-Aware Semantic Segmentation

Multi-view 3D Object Detection Network for Autonomous Driving

PVANet: Lightweight Deep Neural Networks for Real-time Object Detection

Aggregated Residual Transformations for Deep Neural Networks

Weakly Supervised Top-down Salient Object Detection

Deeply Supervised Salient Object Detection with Short Connections

Learning to detect and localize many objects from few examples

Multispectral Deep Neural Networks for Pedestrian Detection

Multiple Representations-Based Face Sketch–Photo Synthesis

StuffNet: Using ‘Stuff’ to Improve Object Detection

Fused DNN: A Deep Neural Network Fusion Approach to Fast and Robust Pedestrian Detection

Gated Bi-directional CNN for Object Detection

A multi-scale cascade fully convolutional network face detector

Visual Saliency Detection Based on Multiscale Deep CNN Features

A deep multi-level network for saliency prediction

Reduced memory region based deep Convolutional Neural Network detection

Edge Preserving and Multi-Scale Contextual Neural Network for Salient Object Detection

Support Vector Machine

Multistage Object Detection With Group Recursive Learning

Deeply-Supervised Recurrent Convolutional Neural Network for Saliency Detection

Saliency Detection via Combining Region-Level and Pixel-Level Predictions with CNNs

Bootstrapping Face Detection with Hard Negative Examples

UnitBox: An Advanced Object Detection Network

A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection

Is Faster R-CNN Doing Well for Pedestrian Detection?

Supervised Transformer Network for Efficient Face Detection

Spatio-Temporal Saliency Networks for Dynamic Saliency Prediction

Two-Stream Convolutional Networks for Dynamic Saliency Prediction

Efficient and robust Pedestrian Detection using Deep Learning for Human-Aware Navigation

A Real-Time Pedestrian Detector using Deep Learning for Human-Aware Navigation

Joint Training of Cascaded CNN for Face Detection

Real-Time Salient Object Detection with a Minimum Spanning Tree

Visual-Inertial-Semantic Scene Representation for 3D Object Detection

Face Detection with the Faster R-CNN

Face Detection with End-to-End Integration of a ConvNet and a 3D Model

Exploit All the Layers: Fast and Accurate CNN Object Detector with Scale Dependent Pooling and Cascaded Rejection Classifiers

R-FCN: Object Detection via Region-based Fully Convolutional Networks

Deep Saliency with Encoded Low Level Distance Map and High Level Features

Subcategory-Aware Convolutional Neural Networks for Object Proposals and Detection

Training Region-Based Object Detectors with Online Hard Example Mining

Recurrent Attentional Networks for Saliency Detection

CRAFT Objects from Images

Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks

A MultiPath Network for Object Detection