Journal of Computer and Communications, 2022, 10, 115-132

https://www.scirp.org/journal/jcc
ISSN Online: 2327-5227
ISSN Print: 2327-5219

Overview of Object Detection Algorithms Using

Convolutional Neural Networks

Junsong Ren, Yi Wang*

School of Computer Science and Engineering, Sichuan University of Science and Engineering, Yibing, China

How to cite this paper: Ren, J.S. and Abstract

Wang, Y. (2022) Overview of Object Detec-
tion Algorithms Using Convolutional Neural In today’s world, computer vision technology has become a very important
Networks. Journal of Computer and Com- direction in the field of Internet applications. As one of the basic problems of
munications, 10, 115-132. computer vision, object detection has become the basis of many vision tasks.
https://doi.org/10.4236/jcc.2022.101006
Whether we need to realize the interaction between images and text or recog-
Received: December 6, 2021 nize fine categories, it provides reliable information. This article reviews the
Accepted: January 26, 2022 development of object detection networks. Starting from RCNN, we intro-
Published: January 29, 2022
duce object detection based on candidate regions, including Fast R-CNN,
Copyright © 2022 by author(s) and
Faster R-CNN, etc.; and then start to introduce single-shot networks includ-
Scientific Research Publishing Inc. ing YOLO, SSD, and Retina Net, etc. Detectors are the most excellent me-
This work is licensed under the Creative thods at present. By reviewing the current research status of object detection
Commons Attribution International
networks, it provides suggestions for the further development trend and re-
License (CC BY 4.0).
http://creativecommons.org/licenses/by/4.0/
search of object detection.
Open Access
Keywords
Deep Learning, Convolutional Neural Network, Object Detection,
Computer Vision

1. Introduction
Convolutional Neural Network [1] (CNN) has made great progress in recent
years and is a very bright pearl in the booming deep neural network treasure
house. And computer vision technology allows artificial intelligence to have the
ability of visual perception and understanding. In recent years, thanks to the
improvement of computer hardware performance and the creation of large-scale
image annotation data sets, computer vision algorithms based on deep learning
have achieved great success in classic computer vision tasks such as image classi-
fication, object detection, and image segmentation.
At present, object detection has not only received a lot of research in acade-

DOI: 10.4236/jcc.2022.101006 Jan. 29, 2022 115 Journal of Computer and Communications
J. S. Ren, Y. Wang

mia, but also has been widely used in real life, such as video fire detection [2],
unmanned driving [3], security monitoring [4], and UAV scene analysis [5]. At
present, object detection algorithms are mainly divided into two types, tradi-
tional object detection algorithms based on image processing and object detec-
tion algorithms based on convolutional neural networks. In 2014, Girshick et al.
proposed R-CNN [6] on this basis. For the first time, convolutional neural net-
works were applied to object detection, and the detection accuracy was im-
proved by nearly 30% compared with traditional detection algorithms, which
caused a great response. From the current academic research and practical ap-
plication, the object detection algorithm based on the convolutional neural net-
work has higher accuracy and shorter test time than the traditional method, and
it has almost completely replaced the traditional algorithm.

2. Convolutional Neural Network

A common CNN example is displayed in Figure 1.
Convolutional neural networks are mainly composed of these types of layers:
input layer, convolutional layer, ReLU layer, pooling layer, and fully connected
layer (the fully connected layer is the same as the conventional neural network).
By superimposing these layers, a complete convolutional neural network can be
constructed. In practical applications, the convolutional layer and the ReLU
layer are often collectively referred to as the convolutional layer, so the convolu-
tional layer also passes through the activation function after the convolution op-
eration. Specifically, when the convolutional layer and the fully connected layer
perform transformation operations on the input, not only the activation func-
tion will be used, but also many parameters, namely the weight w and the devia-
tion b of the neuron; and The ReLU layer and the pooling layer perform a fixed
function operation. The parameters in the convolutional layer and the fully
connected layer will be trained as the gradient drops so that the classification
score calculated by the convolutional neural network can match the label of each
image in the training set.
Convolutional neural networks have the concepts of local receptive fields [7],
sparse weights, and parameter sharing. These three concepts make convolutional
neural networks have a certain translation and scale invariance compared with
other neural networks [8], and are more suitable for image data learning.

Figure 1. Architecture of CNN framework.

DOI: 10.4236/jcc.2022.101006 116 Journal of Computer and Communications

 J. S. Ren, Y. Wang

2.1. Convolutional Layer

The convolutional layer is the core layer to construct a convolutional neural
network, which generates most of the calculations in the network. Note that the
amount of calculation is not the number of parameters. The convolution opera-
tion can effectively reduce the training complexity of the network model and
reduce the network connection and parameter weights, which makes it easier to
train than a fully connected network of the same scale. Common convolution
operations are as follows: ordinary convolution [9], transposed convolution [10],
hole convolution [11], and depth separable convolution [12].
Ordinary convolution is a process in which the convolution kernel is used to
slide on the image, and after a series of matrix operations, the process of calcu-
lating the gray value of all image pixels is finally completed. Transposed Convo-
lution is the convolution method from low-dimensional feature mapping to
high-dimensional feature mapping, which is the opposite of ordinary convolu-
tion. It is widely used in semantic analysis [13], image recognition [14], and
other fields. Fractionally-Strided Convolution realizes the sampling operation of
input features by reducing the step size of transposed convolution and improv-
ing the feature dimension. Transposed Convolution, also known as dilated con-
volution, is a convolution method that does not increase the number of parame-
ters while increasing the receptive field [7] of the unit. Depthwise Separable
Convolution is used in the lightweight network model MobileNets [12], where a
single filter is applied to each input channel by depthwise convolution, and then
pointwise convolution is used to combine the outputs of different depth convo-
lutions. Depth separable convolution realizes the separation of channels and re-
gions in ordinary convolution operations. This decomposition process can
greatly reduce the amount of calculation and the size of the model.

2.2. Activation Layer

Activation Function is a function added to artificial neural networks to help the
network learn complex patterns in data. Similar to the neuron-based model in
the human brain, the activation function ultimately determines the content to be
emitted to the next neuron.
Common activation functions include Rectified Linear Unit (ReLU) [15],
Randomized LeakyReLU (RReLU) [16], Exponential Linear Units (ELU) [17]
and so on. The linear rectification function ReLU is one of the most significant
unsaturated activation functions. As shown in Figure 2, its mathematical ex-
pression is as follows:
f ( x ) = max ( 0, x ) (1)

2.3. Pooling Layer

The pooling layer was first seen in the LeNet [18] article, called Subsample, and
named after the publication of the AlexNet [15] paper. It is one of the commonly
used components in current convolutional neural networks. The pooling layer is

DOI: 10.4236/jcc.2022.101006 117 Journal of Computer and Communications

J. S. Ren, Y. Wang

Figure 2. ReLU function image.

sandwiched between successive convolutional layers to compress the amount of

data and parameters to reduce overfitting. If the input is an image, then the main
function of the pooling layer is to compress the image.
The pooling layer can effectively reduce the size of the matrix, that is, it can
perform collective statistical operations on the special diagnosis at different po-
sitions in the local area of the image, thereby alleviating the excessive sensitivity
of the convolutional layer to the image position, reducing the parameters in the
final fully connected layer, and speeding up calculation speed. The common op-
erations of the pooling layer include the following: max-pooling [19], average
pooling [20], Spatial Pyramid Pooling [18], etc.

3. Classification of Object Detection Algorithms

In recent years, with the development of deep learning, object detection algo-
rithms have made great breakthroughs. The current popular object detection
methods can be divided into two categories. One is the R-CNN algorithm based
on Region Proposal, such as R-CNN, Fast R-CNN [21], Faster R-CNN [22], etc.
They are two-stage and require the First use of heuristic methods for example
Selective search [23], or CNN network to generate Region Proposal [22] and
then perform classification and regression on Region Proposal. The other is
one-stage algorithms such as Yolo [24] and SSD [25], which only use a CNN
network to directly predict the categories and positions of different targets. The
two-stage object detection algorithm needs to perform region extraction opera-
tions, first use the CNN backbone network to extract image features, then find
possible candidate regions from the feature map, and finally perform sliding
window operations on the candidate regions to further determine the target cat-
egory and location information.
The one-stage object detection algorithm does not extract candidate regions

DOI: 10.4236/jcc.2022.101006 118 Journal of Computer and Communications

 J. S. Ren, Y. Wang

through the intermediate layer, but performs feature extraction, target classifica-
tion, and position regression in the entire convolutional network, and then ob-
tains the target position and category. The recognition accuracy is slightly weak-
er than that of the two-stage object detection algorithm. Under the premise, the
speed has been greatly improved. The development process of one-stage and
two-stage algorithms is shown in Figure 3 and Figure 4, respectively.

4. Common Object Detection Network Model

4.1. R-CNN
In 2014, Ross Girshick proposed R-CNN [6], which uses a selective search algo-
rithm to replace the sliding window, which solves the problem of window re-
dundancy and reduces the time complexity of the algorithm. At the same time,
the traditional hand-made feature extraction part is replaced with a convolu-
tional neural network, which can more effectively extract the features of the im-
age and improve the network’s anti-interference ability.
RCNN first selects possible object frames from a set of object candidate frames
through Selective Search algorithm, and then resizes the images in these selected
object frames to a fixed size image, and feeds them to CNN The model (a CNN
model trained on the ImageNet data set, such as AlexNet) extracts features, and
finally sends the extracted features to the classifier to predict whether the image
in the object frame has the target to be detected, and uses the regression to fur-
ther predict Which category does the detection target belong to.
The performance of the R-CNN model has been greatly improved compared
to traditional object detection algorithms, but there are also many problems. For
example, R-CNN generates about 2000 candidate regions, and feature extraction
takes too much time; convolutional neural networks require fixed-size input,
and image cropping or stretching will cause loss of image information; training
speed is slow, not only training image classification The network also needs to
train the SVM [26] classifier and regressor. The structure of R-CNN network is
shown in Figure 5.

4.2. SPPNet
In 2015, SPPNet [18] was published on IEEE. In R-CNN, to generate a vector of
equal dimensions for all candidate regions, the candidate regions are forcibly
scaled, which will destroy the proportional relationship of the image, which is

Figure 3. The development history of the two-stage object detection network framework.

Figure 4. The development history of the one-stage object detection network framework.

DOI: 10.4236/jcc.2022.101006 119 Journal of Computer and Communications

J. S. Ren, Y. Wang

Figure 5. The architecture of the R-CNN framework.

not good for feature extraction, and this extraction process is quite time-consuming,
so SPPNet is optimized here, using spatial pyramid pooling.
The spatial pyramid pooling layer in the figure below is the core of SPPNet,
and its main purpose is to generate a fixed size output for any size input. The
idea is to first divide a feature map of any size into 16, 4, or 1 blocks, and then
pool the maximum on each block. The pooled features are spliced to obtain a
fixed-dimensional output to meet the needs of the fully connected layer. Ob-
viously, for images of different sizes, we get vectors of the same size. This is the
advantage of spatial pyramid pooling. The structure of SPPNet network is shown
in Figure 6.

4.3. Fast R-CNN

Fast R-CNN published in 2015, Comparing Fast R-CNN and R-CNN frame-
works, it can be found that there are two main differences: one is that an ROI
pooling layer is added after the last convolutional layer, and the other is that the
loss function uses a multi-task loss function (multi-task loss), The Bounding Box
Regression is directly added to the CNN network for training.
Fast R-CNN uses the CNN network to first extract the features of the entire
image instead of extracting multiple times for each image block. Then, we can
apply the method of creating candidate regions directly to the extracted feature
maps. For example, Fast R-CNN chose the conv5 layer in VGG16 to generate the
mapped feature block of the ROI region on the corresponding feature map and
used it in the object detection task. We use ROI pooling to convert feature tiles
to a fixed size and send them to the fully connected layer for classification and
positioning. Because Fast-RCNN does not repeatedly extract features, it can sig-
nificantly reduce processing time. Fast R-CNN uses Softmax Loss and Smooth
L1 Loss to jointly train classification probability and Bounding box regression in
the training process. The structure of Fast R-CNN network is shown in Figure 7.

DOI: 10.4236/jcc.2022.101006 120 Journal of Computer and Communications

 J. S. Ren, Y. Wang

Figure 6. The architecture of the SPPNet framework.

Figure 7. The architecture of Fast R-CNN framework.

4.4. Faster R-CNN

The problem with Fast R-CNN: There is a bottleneck: selective search to find all
candidate boxes, which is also very time-consuming. To obtain these candidate
frames more efficiently, Faster R-CNN has added a neural network region pro-
posal network RPN (region proposal network) that extracts edges. After the
convolutional neural network is added RPN, the work of finding candidate
frames can be completed, the region proposal network As part of the Faster

DOI: 10.4236/jcc.2022.101006 121 Journal of Computer and Communications

J. S. Ren, Y. Wang

R-CNN model, it is trained together with the entire model. Realizing the inte-
gration of candidate frame extraction into the deep network, RPN can learn how
to generate high-quality proposed regions, reduce the number of proposed re-
gions learned from the data, and still maintain the accuracy of object detection.
Faster R-CNN [22] is another masterpiece of the author Ross Girshick after
Fast R-CNN. It also uses VGG-16 as the backbone of the network. The inference
speed reaches 5fps on the GPU (including the generation of candidate regions),
that is, it can detect every second of Five pictures, the accuracy rate has also been
further improved, and won first place in multiple projects in the 2015 ILSVRC
and COCO competitions. The structure of Faster R-CNN network is shown in
Figure 8.

4.5. Mask R-CNN

HeKaiming launched Mask R-CNN [27] on ICCV in 2017. Mask R-CNN is an
extension of the original Faster-RCNN, adding a branch to use existing detection
to predict the target in parallel. At the same time, this network structure is rela-
tively easy to implement and train, and the speed is relatively fast, and it can be
easily applied to other fields, such as object detection, segmentation, and key
point detection of people.
The image is first extracted through ResNet-FPN for feature extraction; then
through the RPN network to predict Proposal; then RoI Align is used for feature
extraction, then the classification and detection heads, and finally the Mask de-
tection head, that is, each category predicts a Mask.

Figure 8. The architecture of faster R-CNN framework.

DOI: 10.4236/jcc.2022.101006 122 Journal of Computer and Communications

 J. S. Ren, Y. Wang

RoI Pooling extracts fixed-scale features by performing pooling operations on

the feature map by round, but two rounding operations will cause large errors.
The first rounding is when the proposal of RPN is mapped to the feature map
for rounding. The second rounding is the deviation of the coordinates due to the
image downsampling when the feature map is mapped to the original image.
So RoI Align solves this problem in two ways, The first way is to map the
proposal to the fractional part of the feature map position. The second method is
to use bilinear interpolation to process the feature map where the proposal is lo-
cated to avoid inaccurate features caused by the boundary. The structure of
Mask R-CNN network is shown in Figure 9.

4.6. Trident Net

In 2019, Li et al. proposed TridentNet [28], which was the first to propose the
influence of receptive fields on objects of different scales in object detection
tasks, and carried out relevant experimental verifications. Using the parameter
sharing method, three branches are proposed during training, and only one of
the branches is used during testing, to ensure that there will be no additional
parameters and an increase in calculation during inference.
The TridentNet module mainly includes 3 of the same branches, the only dif-
ference is the expansion rate of the expanded convolution. From top to bottom,
the expansion rates are 1, 2, and 3 respectively, which can detect small, medium,
and large targets respectively, which can better realize multi-scale object detec-
tion. The three branches share weights.
The whole model successfully solves the problem of constant scale in object
detection, but it does not improve the detection speed. The structure of Tri-
dentNet network is shown in Figure 10.

4.7. D2Det
In 2020, based on the two-stage method, Cao et al. improved the classification

Figure 9. The architecture of Mask R-CNN framework.

DOI: 10.4236/jcc.2022.101006 123 Journal of Computer and Communications

J. S. Ren, Y. Wang

Figure 10. The architecture of Trident Net framework.

and regression branches to further improve the accuracy of object detection and
instance segmentation. They proposed D2Det [29], a method that can both ac-
curately locate and accurately classify.
For precise positioning, this paper introduces a dense local regression method,
which is used to predict multiple dense box offsets for each target candidate box.
For accurate classification, this paper introduces a discriminative RoI pooling
scheme. For a candidate area, it can sample from different sub-regions, and then
assign adaptive weights during the calculation to obtain discriminative features.
The structure of D2Det network is shown in Figure 11.

4.8. Sparse R-CNN

Most of the previous target detectors are dense detectors, which are based on
dense recommendations (sliding-windows, anchor-boxes, reference-points),
which are present in the image grid or feature map network in advance On the
grid, the scores and offsets of these suggestions are predicted, judged by IOU
(Intersection Ratio), and then filtered by NMS (Non-Maximum Suppression)
[30].
The small part is the dense-sparse detector (Dense-Sparse), which first ex-
tracts relatively few foreground boxes from the dense suggested regions, that is,
regional candidate boxes, and then classifies and regresses the position of each
regional candidate box, from thousands of candidates, were eliminated to a few
prospects.
Sparse R-CNN [31] avoids the manual setting of a large number of hyperpa-
rameters for candidate boxes and many-to-one positive and negative sample al-
location. More importantly, the final prediction result can be directly output
without NMS (Non-Maximum Suppression).

4.9. YOLOv1
YOLO [24] was proposed in 2016 and published in CVPR, the computer vision
conference.

DOI: 10.4236/jcc.2022.101006 124 Journal of Computer and Communications

 J. S. Ren, Y. Wang

Figure 11. The architecture of the D2Det framework.

Unlike the R-CNN series that needs to find the candidate area first, and then
identify the objects in the candidate area, YOLO’s prediction is based on the en-
tire picture, and it will output all detected target information at one time, in-
cluding category and location.
The first step of YOLO is to divide the picture. It divides the picture into gr-
ids, and the size of each grid is equal. The core idea of YOLO is to turn object
detection into a regression problem, using the entire image as the input of the
network, and only going through a neural network to get the location of the
bounding box and its category. Its detection speed is extremely fast, the genera-
lization ability is strong, the speed is provided, and the accuracy is reduced. The
disadvantage is that for small objects, overlapping objects cannot be detected.

4.10. YOLOv2
In 2017, Joseph Redmon and Ali Farhadi made a lot of improvements based on
YOLOv1, and proposed YOLOv2 [32], focusing on solving the shortcomings of
YOLOv1’s recall rate and positioning accuracy.
Compared with YOLOv1, which uses the fully connected layer to directly pre-
dict the coordinates of the Bounding Box, YOLOv2 draws on the idea of Faster
R-CNN and introduces the Anchor mechanism. The K-means clustering method
is used to cluster and calculate a better Anchor template in the training set,
which greatly improves the recall rate of the algorithm. At the same time, com-
bining the fine-grained features of the image, the shallow features are connected
with the deep features, which is helpful for the detection of small-scale targets.
The article proposes a new training method—a joint training algorithm. This
algorithm can mix these two data sets. Use a hierarchical view to classify objects,
and use a huge amount of classification data set data to expand the detection da-
ta set, thereby mixing two different data sets.

4.11. YOLOv3
In 2018, Redmon made some improvements based on YOLOv2. The feature ex-
traction part uses the darknet-53 network structure to replace the original dark-

DOI: 10.4236/jcc.2022.101006 125 Journal of Computer and Communications

J. S. Ren, Y. Wang

net-19 and uses the feature pyramid network structure to achieve multi-scale
detection. The classification method uses logistic regression instead of softmax
to ensure the accuracy of object detection while taking into account real-time
performance.
YOLOv3’s prior detection system reuses the classifier or locator to perform
detection tasks. They apply the model to multiple locations and scales of the im-
age. Those areas with higher scores can be regarded as the test results. In addi-
tion, compared to other object detection methods, they use a completely differ-
ent method. They apply a single neural network to the entire image. The net-
work divides the image into different regions and predicts the bounding box and
probability of each region. These bounding boxes are weighted by the predicted
probability. The model has some advantages over classifier-based systems. It
looks at the entire image during the test, so its prediction uses the global infor-
mation in the image. Unlike R-CNN, which requires thousands of single target
images, it makes predictions through a single network evaluation. This makes
YOLOv3 [33] very fast, generally, it is 1000 times faster than R-CNN and 100
times faster than Fast R-CNN.

4.12. YOLOv4
In 2020, Bochkovskiy and others launched YOLOv4 [33]. YOLOv4 conducted a
lot of tests on some commonly used Tricks in deep learning and finally selected
these useful Tricks: WRC, CSP, CmBN, SAT, Mish activation, Mosaic data aug-
mentation, CmBN, DropBlock regularization, and CIoU loss. YOLOv4 adds
these practical skills based on traditional YOLO to achieve the best trade-off
between detection speed and accuracy.

4.13. SSD
The full name of the SSD [25] algorithm is Single Shot MultiBox Detector. Sin-
gle-shot indicates that the SSD algorithm is a one-stage method, and MultiBox
indicates that the SSD is a multi-frame prediction.
Compared with Yolo, SSD uses CNN to directly perform detection instead of
performing detection after the fully connected layer as Yolo does. In fact, the di-
rect detection of convolution is only one of the differences between SSD and
Yolo. There are also two important changes. One is that the SSD extracts feature
maps of different scales for detection. Feature maps) can be used to detect small
objects, while small-scale feature maps (the later feature maps) can be used to
detect large objects; the second is that SSD uses prior boxes with different scales
and aspect ratios (Prior boxes, Default boxes), Called Anchors in Faster R-CNN).
The disadvantage of the Yolo algorithm is that it is difficult to detect small tar-
gets and the positioning is not accurate, but these important improvements ena-
ble SSD to overcome these shortcomings to a certain extent.
SSD uses VGG16 as the basic model, and then adds a new convolutional layer
based on VGG16 to obtain more feature maps for detection.

DOI: 10.4236/jcc.2022.101006 126 Journal of Computer and Communications

 J. S. Ren, Y. Wang

4.14. RetinaNet
In 2017, Lin et al. proposed RetinaNet [34]. They believe that the one-stage me-
thod is fast but not as accurate as the two-stage because the positive and negative
samples are not balanced.
The one-stage detector designed a new loss, focal loss for the obstacle problem
of category imbalance during the training process, and the cross-entropy error of
the regression task was changed to focal loss.
Focal loss is a cross entropy loss that can be dynamically zoomed. When the
confidence of the correct category increases, the zoom factor attenuates to 0. The
zoom factor can automatically reduce the weight of the loss contributed by easy
examples during training so that the model pays attention to hard examples.
FPN serves as the Backbone. It adds a top-down path and a lateral path to the
ResNet [35] network and builds a rich, multi-scale feature pyramid from the
single resolution of the picture. The features of each layer of the pyramid are
used to detect targets of different sizes. The structure of RetinaNet network is
shown in Figure 12.

4.15. CornerNet
In 2018, Hei et al. published CornerNet [36] on ECCV2018. They proposed to
solve the object detection problem as a key point detection problem, that is, to
obtain the prediction frame by detecting the two key points of the upper left
corner and the lower right corner of the target frame. Therefore, there is no
concept of anchor in the CornerNet algorithm. This approach is used in object
detection. The field is relatively innovative and can achieve good results. The
training of the entire detection network is started from scratch and is not based
on a pre-trained classification model. This allows users to freely design a feature
extraction network without being restricted by the pre-training model. Corner-
Net also proposed a new pooling method: corner pooling. The structure of Cor-
nerNet network is shown in Figure 13.

4.16. CenterNet
CenterNet [37], it can be seen from the name of the algorithm that this algorithm

Figure 12. The architecture of RetinaNet framework.

DOI: 10.4236/jcc.2022.101006 127 Journal of Computer and Communications

J. S. Ren, Y. Wang

Figure 13. The architecture of CornerNet framework.

is to predict the center point of the target, instead of the two corner points in
CornerNet; CenterNet uses a heat map to achieve this, introducing the Gaussian
distribution area of the predicted points Calculate the true predicted value. At
the same time, the heat map output by the network will first be normalized to 0
to 1 through the sigmoid function and then transferred to the loss function.
CenterNet does not include operations such as corner pooling, because the
probability of the center point of the target frame falling on the target is rela-
tively large, so the conventional pooling operation can extract effective features.
CenterNet also uses the same offset prediction as CornerNet, which represents
the coordinate error caused by the rounding operation when the annotation in-
formation is mapped from the input image to the output feature map, but the
calculation in CornerNet is 2. The offset of the corner point, and CenterNet cal-
culates the offset of the center point.

4.17. EffcientNet
An et al. proposed the EffcientDet [38] algorithm on CVPR 2020. They believe
that the current object detection, either pursues more accurate detection results,
but costs a lot, or is more efficient, but at the expense of accuracy. Therefore, the
paper designs a set of object detection frameworks to adapt to different con-
straints, while satisfying high precision and high efficiency. They mainly pro-
posed BiFPN and compound scale methods.
BiFPN is an improvement based on FPN. The original FPN module adds
edges to add contextual information and multiplies each edge by corresponding
Weights. It allows simple and fast multi-scale feature fusion, secondly, the com-
pound scale method can uniformly scale the resolution, depth and width, feature
network, and box/class prediction network of all backbones.

4.18. CentripetalNet
CentripetalNet [39] published in CVPR 2020 uses centripetal displacement to
pair corner points in the same instance. CentripetalNet predicts the position and
centripetal shift of corner points and matches the aligned corner points as a re-
sult of the shift.
Combining location information, CentripetalNet matches corner points more
accurately than traditional embedding methods. corner pooling extracts the in-
formation in the bounding box to the boundary. To make the information at the

DOI: 10.4236/jcc.2022.101006 128 Journal of Computer and Communications

 J. S. Ren, Y. Wang

corner points more sensitive, CentripetalNet designed a corner-star deformable

convolution network for feature adaptation. In addition, CentripetalNet also ex-
plores instance segmentation on anchorless detectors by installing a mask pre-
diction module on the centripetal network.

5. Conclusions
In recent years, with the rapid development of deep learning technology, uni-
versal object detection technology has developed rapidly and made break-
throughs. However, there is still a huge gap between the efficiency and speed of
the detection model and the humanized performance. Existing research methods
show that: based on depth The problems to be solved and future research trends
of the learned general object detection technology mainly include:
1) Unsupervised object detection: Automated labeling technology is exciting
and promising. Unsupervised object detection can eliminate manual labeling.
2) To study a detection method that can have the advantages of both Tow
Stage and One Stage models at the same time.
3) Design an efficient feature extraction network.
4) GAN object detector: We know that deep learning object detectors usually
require a lot of data for training. In contrast, the GAN target detector is an im-
portant structure for producing false images. The combination of real scenes and
GAN simulation data helps the detector to be more robust and general.
5) Multi-domain object detection: a general object detector is mainly devel-
oped, which can detect multi-domain objects without prior knowledge.

Funding
Supported by the scientific research project of the Sichuan University of Science
and Engineering (ZHZJ19-01).

Conflicts of Interest
The authors declare no conflicts of interest regarding the publication of this pa-
per.

References
[1] Krizhevsky, A., Sutskever, I. and Hinton, G.E. (2017) ImageNet classification with
Deep Convolutional Neural Networks. Communications of the ACM, 60, 84-90.
https://doi.org/10.1145/3065386
[2] Kim, B. and Lee, J. (2019) A Video-Based Fire Detection Using Deep Learning
Models. Applied Sciences, 9, Article No. 2862. https://doi.org/10.3390/app9142862
[3] Li, P., Chen, X. and Shen, S. (2019) Stereo R-CNN Based 3D Object Detection for
Autonomous Driving. Proceedings of the IEEE/CVF Conference on Computer Vi-
sion and Pattern Recognition, Long Beach, 15-20 June 2019, 7636-7644.
https://doi.org/10.1109/CVPR.2019.00783
[4] Zhang, X., Yi, W.-J. and Saniie, J. (2019) Home Surveillance System Using Com-
puter Vision and Convolutional Neural Network. 2019 IEEE International Confe-

DOI: 10.4236/jcc.2022.101006 129 Journal of Computer and Communications

J. S. Ren, Y. Wang

rence on Electro Information Technology (EIT), Brookings, 20-22 May 2019,

266-270. https://doi.org/10.1109/EIT.2019.8833773
[5] Zhang, R., Shao, Z., Huang, X., Wang, J. and Li, D. (2020) Object Detection in UAV
Images via Global Density Fused Convolutional Network. Remote Sensing, 12, Ar-
ticle No. 3140. https://doi.org/10.3390/rs12193140
[6] Girshick, R., Donahue, J., Darrell, T. and Malik, J. (2013) Rich Feature Hierarchies
for Accurate Object Detection and Semantic Segmentation. 2014 IEEE Conference
on Computer Vision and Pattern Recognition, Columbus, 23-28 June 2014, 580-587.
https://doi.org/10.1109/CVPR.2014.81
[7] Fukushima, K. and Miyake, S. (1982) Neocognitron: A New Algorithm for Pattern
Recognition Tolerant of Deformations and Shifts in Position. Pattern recognition,
15, 455-469. https://doi.org/10.1016/0031-3203(82)90024-3
[8] Singh, B. and Davis, L.S. (2018) An Analysis of Scale Invariance in Object Detection
Snip. Proceedings of the IEEE Conference on Computer Vision and Pattern Recog-
nition, Salt Lake City, 18-23 June 2018, 3578-3587.
https://doi.org/10.1109/CVPR.2018.00377
[9] Ngiam, J., Chen, Z., Chia, D., Koh, P., Le, Q. and Ng, A. (2010) Tiled Convolutional
Neural Networks. Advances in Neural Information Processing Systems, 23.
[10] Zeiler, M.D. and Fergus, R. (2014) Visualizing and Understanding Convolutional
Networks. European Conference on Computer Vision, Zurich, 6-12 September,
818-833. https://doi.org/10.1007/978-3-319-10590-1_53
[11] Yu, F. and Koltun, V. (2015) Multi-Scale Context Aggregation by Dilated Convolu-
tions. https://arxiv.org/abs/1511.07122
[12] Howard, A.G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., et al.
(2017) Mobilenets: Efficient Convolutional Neural Networks for Mobile Vision Ap-
plications. https://arxiv.org/abs/1704.04861
[13] Noh, H., Hong, S. and Han, B. (2015) Learning Deconvolution Network for Seman-
tic Segmentation. Proceedings of the IEEE International Conference on Computer
Vision, Santiago, 7-13 December 2015, 1520-1528.
https://doi.org/10.1109/ICCV.2015.178
[14] Chen, R., Wang, M. and Lai, Y. (2020) Analysis of the Role and Robustness of Ar-
tificial Intelligence in Commodity Image Recognition under Deep Learning Neural
Network. PLoS ONE, 15, Article ID: e0235783.
https://doi.org/10.1371/journal.pone.0235783
[15] Krizhevsky, A., Sutskever, I. and Hinton, G.E. (2012) ImageNet Classification with
Deep Convolutional Neural Networks. Advances in Neural Information Processing
Systems, 25, 1097-1105.
[16] Xu, B., Wang, N., Chen, T. and Li, M. (2015) Empirical Evaluation of Rectified Ac-
tivations in Convolutional Network. https://arxiv.org/abs/1505.00853
[17] Clevert, D.-A., Unterthiner, T. and Hochreiter, S. (2015) Fast and Accurate Deep
Network Learning by Exponential Linear Units (ELUs).
https://arxiv.org/abs/1511.07289
[18] He, K., Zhang, X., Ren, S. and Sun, J. (2015) Spatial Pyramid Pooling in Deep Con-
volutional Networks for Visual Recognition. IEEE Transactions on Pattern Analysis
and Machine Intelligence, 37, 1904-1916.
https://doi.org/10.1109/TPAMI.2015.2389824
[19] Graham, B. (2014) Fractional Max-Pooling. https://arxiv.org/abs/1412.6071
[20] Zhai, S., Wu, H., Kumar, A., Cheng, Y., Lu, Y., Zhang, Z., et al. (2017) S3pool:

DOI: 10.4236/jcc.2022.101006 130 Journal of Computer and Communications

 J. S. Ren, Y. Wang

Pooling with Stochastic Spatial Sampling. Proceedings of the IEEE Conference on

Computer Vision and Pattern Recognition, Honolulu, 21-26 July 2017, 4003-4011.
https://doi.org/10.1109/CVPR.2017.426
[21] Girshick, R. (2015) Fast R-CNN. Proceedings of the IEEE International Conference
on Computer Vision, Santiago, 7-13 December 2015, 1440-1448.
https://doi.org/10.1109/ICCV.2015.169
[22] Ren, S., He, K., Girshick, R. and Sun, J. (2017) Faster R-CNN: Towards Real-Time
Object Detection with Region Proposal Networks. IEEE Transactions on Pattern
Analysis & Machine Intelligence, 39, 1137-1149.
https://doi.org/10.1109/TPAMI.2016.2577031
[23] Uijlings, J.R., van de Sande, K.E.A., Gevers, T. and Smeulders, A.W.M. (2013) Se-
lective Search for Object Recognition. International Journal of Computer Vision,
104, 154-171. https://doi.org/10.1007/s11263-013-0620-5
[24] Redmon, J., Divvala, S., Girshick, R. and Farhadi, A. (2015) You Only Look Once:
Unified, Real-Time Object Detection. 2016 IEEE Conference on Computer Vision
and Pattern Recognition (CVPR), Las Vegas, 27-30 June 2016, 779-788.
https://doi.org/10.1109/CVPR.2016.91
[25] Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y. and Berg, A.C.
(2016) SSD: Single Shot Multibox Detector. European Conference on Computer Vi-
sion, Amsterdam, 8-16 October 2016, 21-37.
https://doi.org/10.1007/978-3-319-46448-0_2
[26] Platt, J. (1998) Sequential Minimal Optimization: A Fast Algorithm for Training
Support Vector Machines.
[27] He, K., Gkioxari, G., Dollár, P. and Girshick, R. (2017) Mask R-CNN. 2017 IEEE
International Conference on Computer Vision (ICCV), Venice, 22-29 October 2017,
2980-298. https://doi.org/10.1109/ICCV.2017.322
[28] Li, Y., Chen, Y., Wang, N. and Zhang, Z.-X. (2019) Scale-Aware Trident Networks
for Object Detection. Proceedings of the IEEE/CVF International Conference on
Computer Vision, Seoul, 27 October-2 November 2019, 6053-6062.
https://doi.org/10.1109/ICCV.2019.00615
[29] Cao, J., Cholakkal, H., Anwer, R.M., Khan, F.S., Pang, Y. and Shao, L. (2020) D2det:
Towards High Quality Object Detection and Instance Segmentation. Proceedings of
the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle,
13-19 June 2020, 11482-11491. https://doi.org/10.1109/CVPR42600.2020.01150
[30] Neubeck, A. and Van Gool, L. (2006) Efficient Non-Maximum Suppression. 18th
International Conference on Pattern Recognition (ICPR’06), Hong Kong, 20-24 August
2006, 850-855. https://doi.org/10.1109/ICPR.2006.479
[31] Sun, P., Zhang, R., Jiang, Y., Kong, T., Xu, C., Zhan, W., et al. (2021) Sparse
R-CNN: End-to-End Object Detection with Learnable Proposals. Proceedings of the
IEEE/CVF Conference on Computer Vision and Pattern Recognition, Nashville,
20-25 June 2021, 14449-14458. https://doi.org/10.1109/CVPR46437.2021.01422
[32] Redmon, J. and Farhadi, A. (2017) YOLO9000: Better, Faster, Stronger. Proceedings
of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu,
21-26 July 2017, 6517-6525. https://doi.org/10.1109/CVPR.2017.690
[33] Bochkovskiy, A., Wang, C.-Y. and Liao, H.-Y.M. (2020) YOLOv4: Optimal Speed
and Accuracy of Object Detection. https://arxiv.org/abs/2004.10934
[34] Lin, T.-Y., Goyal, P., Girshick, R., He, K. and Dollár, P. (2017) Focal Loss for Dense
Object Detection. Proceedings of the IEEE International Conference on Computer
Vision, Venice, 22-29 October 2017, 2999-3007.

DOI: 10.4236/jcc.2022.101006 131 Journal of Computer and Communications

J. S. Ren, Y. Wang

https://doi.org/10.1109/ICCV.2017.324
[35] He, K., Zhang, X., Ren, S. and Sun, J. (2016) Deep Residual Learning for Image
Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern
Recognition, Las Vegas, 27-30 June 2016, 770-778.
https://doi.org/10.1109/CVPR.2016.90
[36] Law, H. and Deng, J. (2018) Cornernet: Detecting Objects as Paired Keypoints.
Proceedings of the European Conference on Computer Vision (ECCV), Munich,
8-14 September 2018, 765-781. https://doi.org/10.1007/978-3-030-01264-9_45
[37] Zhou, X., Wang, D. and Krähenbühl, P. (2019) Objects as Points.
https://arxiv.org/abs/1904.07850
[38] Tan, M. and Le, Q. (2019) EfficientNet: Rethinking Model Scaling for Convolution-
al Neural Networks. International Conference on Machine Learning, Long Beach,
10-15 June 2019, 6105-6114.
[39] Dong, Z., Li, G., Liao, Y., Wang, F., Ren, P. and Qian, C. (2020) Centripetalnet:
Pursuing High-Quality Keypoint Pairs for Object Detection. Proceedings of the
IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, 13-19
June 2020, 10516-10525. https://doi.org/10.1109/CVPR42600.2020.01053

DOI: 10.4236/jcc.2022.101006 132 Journal of Computer and Communications