Review on Object Detection and Recognition in Large Field of View

Tangwei Li; Guanjun Tong; Baoqing Li; Xiaoyang Lu

doi:10.3788/LOP57.120002

Laser & Optoelectronics Progress, Volume. 57, Issue 12, 120002(2020)

Review on Object Detection and Recognition in Large Field of View

Tangwei Li^1,2, Guanjun Tong^1、*, Baoqing Li¹, and Xiaoyang Lu¹

¹Key Laboratory of Science and Technology on Micro-System, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, Shanghai 201800, China

²University of Chinese Academy of Sciences, Beijing 100049, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(12)

Fig. 1. Inconsistent distortion of object

Download full size

Fig. 2. Incomplete and blurring object in LFOV images

Download full size

Fig. 3. Distortion of small object in LFOV images

Download full size

Fig. 4. Asymmetry of object in different LFOV images

Download full size

Fig. 5. Classification of object detection and recognition in large field of view

Download full size

Fig. 6. Flow chart of object detection and recognition based on distortion correction

Download full size

Fig. 7. Road map of object detection and recognition in large field of view

Download full size

Table 1. Comparison of performance of different algorithms

View table

Table 1. Comparison of performance of different algorithms

Paper	Object	Pre-process	Feature extraction	Evaluation	Year
Jeong et al.^[22]	Vehicle	Undistortion using FOV $\begin{matrix} r_{u} = \frac{\tan (r_{d} w)}{2 \tan (\frac{w}{2})} \end{matrix}$	HOG		2016
Silbersteinet al.^[24]	Pedestrian	Camera calibration	AFS	10.3%(avgMRF)	2014
Levi andSilberstein^[25]	Pedestrian	Camera calibration	AFS-Multi-Cue	3.5%(avgMRF)	2015
Bertozziet al.^[26]	Pedestrian	Image un-warping $\begin{matrix} [\begin{matrix} u \\ v \end{matrix}] = [\begin{matrix} c_{u} + atan 2 (x, z) \cdot f_{θ} \\ c_{v} + k_{v} \cdot \frac{y}{\sqrt[]{x^{2} + y^{2}}} \end{matrix}] \end{matrix}$	Soft-Cascadeand ACF	0.35(FPPI is 10^-1),0.59 (FPPI is 10^-2),0.75 (FPPI is 10^-3)	2015
Suhr et al.^[28]	Pedestrian	Image warping $\begin{matrix} \begin{matrix} θ = \arctan (x / f) \\ ϕ = \arctan (y / \sqrt[]{x^{2} + y^{2}}) \\ x' = sθ \\ y' = sln (secϕ + tanϕ) \end{matrix} \end{matrix}$	HOG andTER-based classifier	97.3%(TWA)	2017

Table 2. Comparison of performance of different algorithms

View table

Table 2. Comparison of performance of different algorithms

Paper	Object	Pre-process	Feature extraction	Evaluation	Year
Martinezet al.^[31]	Humanbeings	Transformation $\begin{matrix} \begin{matrix} ϕ = c / R_{med} \\ R_{med} = (R_{\min} + R_{\max} \frac{)}{2.0} \\ c_{0} = c_{1} + (R_{\min} + r) sinϕ \\ r_{0} = r_{1} + (R_{\min} + r) cosϕ \end{matrix} \end{matrix}$	Viola-Jonesclassifier	Time reducedfrom600-700 ms to10-15 ms	2010
Dinget al.^[32]	Motionobject	$\begin{matrix} \begin{matrix} \bar{x} = x + {k_{1} x (x^{2} + y^{2}) + s_{1} (x^{2} + y^{2}) + \\ [p_{1} (3 x^{2} + y^{2}) + 2 p_{1} xy]} \\ \bar{y} = y + {k_{2} x (x^{2} + y^{2}) + s_{2} (x^{2} + y^{2}) + \\ [p_{2} (3 x^{2} + y^{2}) + 2 p_{2} xy]} \end{matrix} \end{matrix}$	Based ontexture andcolor feature	5× faster	2016

Table 3. Comparison of performance of different algorithms

View table

Table 3. Comparison of performance of different algorithms

Paper	Object	Pre-process	Feature extraction	Evaluation	Date
Yoshimiet al.^[36]	Pedestrian	HPM	Faster R-CNN	LAMR: 24.5%;E=48.65%	2017
Cai et al.^[37]	VOC Pascal	Cylindrical unwarpingCorrection	YOLO	Detection rate: 30.89 frame·s^-1Accuracy rate: 72%	2018
Xu et al.^[38]	Face	Spherical projection	LCNN	97.3%(TWA)	2018
Deng et al.^[40]	Fire	Spherical projection	CNN		2017

Table 4. Comparison of performance of different algorithms

View table

Table 4. Comparison of performance of different algorithms

Method	Advantage	Disadvantage
Camera calibration	Higher precision	Complexity
	No restrictions on camera type	Need to know a certain size of calibration object
Non-metric calibration	Simple	Worse stability
	Few parameters	Not suitable for high-precision system

Table 5. Comparison of performance of different algorithms

View table

Table 5. Comparison of performance of different algorithms

Paper	Object	Method	Evaluation	Year
Deng et al.^[42]	20 class	Faster R-CNN	mAP: 68.7%	2017
Coors et al.^[49]	Flying cars	Encode the invariance of geometrictransformation directly into CNN	mAP: 50.18%	2018
Herceg et al.^[62]	Motion object	Corner detection, optical flow	Accuracy: 97.19%	2011
Wu et al.^[65]	Motion object	Moving blob method; PTZ shots image	Accuracy: 92%	2017
Zhang et al.^[67]	Salient object	Co-saliency detection algorithm	Precision: 0.82;recall: 0.75; F₁: 0.81	2017
Cinaroglu et al.^[69]	People	Omnidirectional sliding windowModified HOG+SVM	SVM scores: 2.94	2014
Wang et al.^[70]	People	Template-based	MOTA: 0.85	2017

Tools

Get Citation

Copy Citation Text

Tangwei Li, Guanjun Tong, Baoqing Li, Xiaoyang Lu. Review on Object Detection and Recognition in Large Field of View[J]. Laser & Optoelectronics Progress, 2020, 57(12): 120002

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Reviews

Received: Jul. 19, 2019

Accepted: Oct. 31, 2019

Published Online: Jun. 3, 2020

The Author Email: Guanjun Tong (tongguanjun@mail.sim.ac.cn)

DOI:10.3788/LOP57.120002

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology