SlideShare a Scribd company logo
1 of 1
Ensemble Normalization for Stable Training
𝑆𝑒𝑜𝑢𝑛𝑔-𝐻𝑜 𝐶ℎ𝑜𝑖 * and 𝐽𝑖 𝑊𝑜𝑜𝑛𝑔 𝐶ℎ𝑜𝑖
1. Department of Electronic Information Engineering, Hansung University
2. Department of Trade, Hansung University
1. Introduction
Research on normalization is needed to correct and learn internal covariate shifts.
Excessive internal covariate shifts can result in vanishing or exploding gradients. However,
learning by conventional single normalization can be learned in an unoptimized direction.
In order to guide the learning of different features in the right direction through learning,
we propose an ensemble normalization that learns different features through two
normalizations.
Abstract Normalization studies have been conducted to solve vanishing or exploding gradient for stable training. This phenomenon leads to poor model training. Existing normalization method has used a
single normalization. However, single normalization has limitations in stabilizing learning through the model weight of internal covariate shift prevention. For improvement, we propose new
normalization method, an ensemble method that combines two normalization methods. We analyze the existing normalization method through an ablation study. To verify the proposed method,
We experimented two models and two datasets in semantic segmentation task and two models and a dataset in generation task. As a result, we confirmed that the new normalization method is
effective for improving stable training and semantic segmentation and generation performance.
2. Related Works
Existing normalization studies have been conducted such as batch normalization (BN),
instance normalization (IN), and group normalization (GN). BN normalizes samples in mini-
batch units through using mean and variance of each mini-batch and also add scaling and
shifting factors to the layers. However, BN does not perform well at small batch sizes. IN
normalizes across each channel in each sample. IN performs well on style transfer
regardless of the batch size. GN is similar to IN but normalizes over groups of channels for
each sample. GN performed as well as BN on ImageNet.
3. Ensemble Normalization
Figure 1. Process of Normalization Method, a) Existing normalization method, b)
Ensemble normalization method
Ensemble normalization calculates the addition of batch normalization and existing
normalization. Then apply division 2 to the preceding result. The above calculation
sequence is shown in Figure 1. Input is a group of feature maps convoluted by a
convolutional layer. Output is a group of feature maps normalizaed by a normalization
layer which is one of the conventional methods in a). However in b), the normalization
layer is consist of two methods. To have a clear understanding of the existing methods, We
conducted the ablation study on normalization to conduct an impact analysis on each
existing normalization method before verifying the proposed method. In semantic
segmentation task, we analyzed each normalization in the learning process through Focal
loss, Hinge loss, and Cross entropy. Experimental results show average results from Focal
loss, Hinge loss, and Cross entropy. The experimental method uses VOC and ATR datasets
for FCN and U-Net and uses the average value of three losses and uses the average value
through five iterative experiments. In generation task, we analyzed each normalization by
using GAN and LSGAN models and MNIST dataset.
5. Conclusion
We propose a new normalization which is an ensemble normalization. First, the impact of
existing normalization was normalization was analyzed through ablation studies to
determine the impact of the existing method. We also verified the ensemble normalization
method proposed in this paper. As a result, we confirmed that the proposed method is
effective for stable training on semantic segmentation and generation.
a) b)
FCN
FCNIN
FCNBN
FCNGN
Figure 2. Four feature maps in the second layer of the model, each with BN, IN, and GN
applied. a) existing normalizations, b) the proposed normalization.
As shown in figure 2, a comparison is made between the conventional normalization and
the proposed method. In the case of existing normalization, each feature map is not
different from each other, whereas in the feature map of the proposed method, different
feature maps can be confirmed by looking at BIN (combined BN with IN) and BGN
(combined BN with GN). This means that many different features have been trained,
which leads to performance gains. In the case of BBN (combined BN with BN), on the
other hand, three of the four feature maps show little difference. This means that rich
features are not trained, and as shown as in table2, they do not perform relatively well.
GAN
GANBN
GANGN
GANBBN
GANIN GANBIN
GANBGN
LSGAN
LSGANBN
LSGANGN
LSGANBBN
LSGANIN LSGANBIN
LSGANBGN
Figure 2. Four generated images of GAN and LSGAN with conventional normalization
methods or ensemble normalization methods.
4. Experimental Results
Table1. Quantitative comparison of two segmentation models(FCN and U-Net) using the
VOC dataset and two generation models(GAN and LSGAN) using the MNIST dataset.
Table 1 shows experimental results using two datasets for the existing methods and the
proposed method and overall performance is improved. It is because they learn
complementary to each other by catching the characteristics that cannot be captured by
existing normalization methods through other normalization methods. With the addition of
instance normalization, batch normalization, and group normalization, the loss each
changes were +2.3%, +0.4%, and -4.2% for the FCN model and in the U-Net model, there
were -42.9%, -20.3%, and -54.2%, respectively.
Also, overall structure similarity(SSIM) values have been improved in the generator
models. Through this experiment, we confirmed that the proposed method works well for
the generator models.
As shown as in figure 2, it is confirmed that By applying the proposed normalization
method, the image is clearer than when applying the conventional normalization method.

More Related Content

Similar to Ensemble normalization for stable training

An Improved Method for Reconstruction and Enhancing Dark Images
An Improved Method for Reconstruction and Enhancing Dark ImagesAn Improved Method for Reconstruction and Enhancing Dark Images
An Improved Method for Reconstruction and Enhancing Dark Images
Pavan A C
 
Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...
Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...
Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...
ijsrd.com
 
(DL Hacks輪読) How transferable are features in deep neural networks?
(DL Hacks輪読) How transferable are features in deep neural networks?(DL Hacks輪読) How transferable are features in deep neural networks?
(DL Hacks輪読) How transferable are features in deep neural networks?
Masahiro Suzuki
 
Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...
ijma
 
A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...
A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...
A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...
CSCJournals
 

Similar to Ensemble normalization for stable training (20)

[ICCV 21] Influence-Balanced Loss for Imbalanced Visual Classification
[ICCV 21] Influence-Balanced Loss for Imbalanced Visual Classification[ICCV 21] Influence-Balanced Loss for Imbalanced Visual Classification
[ICCV 21] Influence-Balanced Loss for Imbalanced Visual Classification
 
International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)International Journal of Engineering Research and Development (IJERD)
International Journal of Engineering Research and Development (IJERD)
 
Batch_Normalization.pptx
Batch_Normalization.pptxBatch_Normalization.pptx
Batch_Normalization.pptx
 
final paper1
final paper1final paper1
final paper1
 
An Improved Method for Reconstruction and Enhancing Dark Images
An Improved Method for Reconstruction and Enhancing Dark ImagesAn Improved Method for Reconstruction and Enhancing Dark Images
An Improved Method for Reconstruction and Enhancing Dark Images
 
Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...
Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...
Introduction to Wavelet Transform and Two Stage Image DE noising Using Princi...
 
V.KARTHIKEYAN PUBLISHED ARTICLE
V.KARTHIKEYAN PUBLISHED ARTICLEV.KARTHIKEYAN PUBLISHED ARTICLE
V.KARTHIKEYAN PUBLISHED ARTICLE
 
PR422_hyper-deep ensembles.pdf
PR422_hyper-deep ensembles.pdfPR422_hyper-deep ensembles.pdf
PR422_hyper-deep ensembles.pdf
 
Improvement of Image Deblurring Through Different Methods
Improvement of Image Deblurring Through Different MethodsImprovement of Image Deblurring Through Different Methods
Improvement of Image Deblurring Through Different Methods
 
Neural Networks with Complex Sample Data
Neural Networks with Complex Sample DataNeural Networks with Complex Sample Data
Neural Networks with Complex Sample Data
 
(DL Hacks輪読) How transferable are features in deep neural networks?
(DL Hacks輪読) How transferable are features in deep neural networks?(DL Hacks輪読) How transferable are features in deep neural networks?
(DL Hacks輪読) How transferable are features in deep neural networks?
 
Sequence alignment
Sequence alignmentSequence alignment
Sequence alignment
 
Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...
 
Multiple Ant Colony Optimizations for Stereo Matching
Multiple Ant Colony Optimizations for Stereo MatchingMultiple Ant Colony Optimizations for Stereo Matching
Multiple Ant Colony Optimizations for Stereo Matching
 
Review on Image Enhancement in Spatial Domain
Review on Image Enhancement in Spatial DomainReview on Image Enhancement in Spatial Domain
Review on Image Enhancement in Spatial Domain
 
4
44
4
 
H017416670
H017416670H017416670
H017416670
 
A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...
A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...
A Study of Total-Variation Based Noise-Reduction Algorithms For Low-Dose Cone...
 
C04302025030
C04302025030C04302025030
C04302025030
 
An Efficient APOA Techniques For Generalized Residual Vector Quantization Bas...
An Efficient APOA Techniques For Generalized Residual Vector Quantization Bas...An Efficient APOA Techniques For Generalized Residual Vector Quantization Bas...
An Efficient APOA Techniques For Generalized Residual Vector Quantization Bas...
 

More from Seoung-Ho Choi

More from Seoung-Ho Choi (20)

Seoung-Ho Choi Introduction to medical deep learning
Seoung-Ho Choi Introduction to medical deep learningSeoung-Ho Choi Introduction to medical deep learning
Seoung-Ho Choi Introduction to medical deep learning
 
Seungho Choi Introduction to deep learning solutions
Seungho Choi Introduction to deep learning solutionsSeungho Choi Introduction to deep learning solutions
Seungho Choi Introduction to deep learning solutions
 
To classify Alzheimer’s Disease from 3D structural MRI data
To classify Alzheimer’s Disease from 3D structural MRI dataTo classify Alzheimer’s Disease from 3D structural MRI data
To classify Alzheimer’s Disease from 3D structural MRI data
 
Middle school winter science garden participation certificate
Middle school winter science garden participation certificateMiddle school winter science garden participation certificate
Middle school winter science garden participation certificate
 
Elementary school winter model aircraft school certificate
Elementary school winter model aircraft school certificateElementary school winter model aircraft school certificate
Elementary school winter model aircraft school certificate
 
Elementary school youth science exploration contest silver prize
Elementary school youth science exploration contest silver prizeElementary school youth science exploration contest silver prize
Elementary school youth science exploration contest silver prize
 
Elementary school youth science contest silver prize
Elementary school youth science contest silver prizeElementary school youth science contest silver prize
Elementary school youth science contest silver prize
 
Middle school creativity problem solving ability contest encouragement prize
Middle school creativity problem solving ability contest encouragement prizeMiddle school creativity problem solving ability contest encouragement prize
Middle school creativity problem solving ability contest encouragement prize
 
Middle school youth science exploration contest bronze prize
Middle school youth science exploration contest bronze prizeMiddle school youth science exploration contest bronze prize
Middle school youth science exploration contest bronze prize
 
Elementary school minister of science and technology award
Elementary school minister of science and technology awardElementary school minister of science and technology award
Elementary school minister of science and technology award
 
Elementary school completion certificate Jung-gu education center for the gifted
Elementary school completion certificate Jung-gu education center for the giftedElementary school completion certificate Jung-gu education center for the gifted
Elementary school completion certificate Jung-gu education center for the gifted
 
Encouragement award in Korean Information Science Society for Undergraduate S...
Encouragement award in Korean Information Science Society for Undergraduate S...Encouragement award in Korean Information Science Society for Undergraduate S...
Encouragement award in Korean Information Science Society for Undergraduate S...
 
Best paper in Korean Communication Society
Best paper in  Korean Communication SocietyBest paper in  Korean Communication Society
Best paper in Korean Communication Society
 
PS(Personal Statement) Korean
PS(Personal Statement) KoreanPS(Personal Statement) Korean
PS(Personal Statement) Korean
 
PS(Personal Statement) English
PS(Personal Statement) EnglishPS(Personal Statement) English
PS(Personal Statement) English
 
A Study on the Importance of Adaptive Seed Value Exploration
A Study on the Importance of Adaptive Seed Value Exploration A Study on the Importance of Adaptive Seed Value Exploration
A Study on the Importance of Adaptive Seed Value Exploration
 
Bi-activation Function : an Enhanced Version of an Activation Function in C...
Bi-activation Function : an Enhanced Version of an Activation Function in C...Bi-activation Function : an Enhanced Version of an Activation Function in C...
Bi-activation Function : an Enhanced Version of an Activation Function in C...
 
Visualization Techniques for Outlier data
Visualization Techniques for Outlier data Visualization Techniques for Outlier data
Visualization Techniques for Outlier data
 
Gpt1 and 2 model review
Gpt1 and 2 model reviewGpt1 and 2 model review
Gpt1 and 2 model review
 
Unsupervised learning for real-world super-resolution review presentation
Unsupervised learning for real-world super-resolution review presentationUnsupervised learning for real-world super-resolution review presentation
Unsupervised learning for real-world super-resolution review presentation
 

Recently uploaded

TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...
TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...
TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...
TrustArc
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
panagenda
 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Safe Software
 

Recently uploaded (20)

Design Guidelines for Passkeys 2024.pptx
Design Guidelines for Passkeys 2024.pptxDesign Guidelines for Passkeys 2024.pptx
Design Guidelines for Passkeys 2024.pptx
 
TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...
TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...
TrustArc Webinar - Unified Trust Center for Privacy, Security, Compliance, an...
 
Understanding the FAA Part 107 License ..
Understanding the FAA Part 107 License ..Understanding the FAA Part 107 License ..
Understanding the FAA Part 107 License ..
 
Stronger Together: Developing an Organizational Strategy for Accessible Desig...
Stronger Together: Developing an Organizational Strategy for Accessible Desig...Stronger Together: Developing an Organizational Strategy for Accessible Desig...
Stronger Together: Developing an Organizational Strategy for Accessible Desig...
 
Introduction to use of FHIR Documents in ABDM
Introduction to use of FHIR Documents in ABDMIntroduction to use of FHIR Documents in ABDM
Introduction to use of FHIR Documents in ABDM
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
 
Less Is More: Utilizing Ballerina to Architect a Cloud Data Platform
Less Is More: Utilizing Ballerina to Architect a Cloud Data PlatformLess Is More: Utilizing Ballerina to Architect a Cloud Data Platform
Less Is More: Utilizing Ballerina to Architect a Cloud Data Platform
 
Platformless Horizons for Digital Adaptability
Platformless Horizons for Digital AdaptabilityPlatformless Horizons for Digital Adaptability
Platformless Horizons for Digital Adaptability
 
ADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptxADP Passwordless Journey Case Study.pptx
ADP Passwordless Journey Case Study.pptx
 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
 
Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...
Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...
Web Form Automation for Bonterra Impact Management (fka Social Solutions Apri...
 
Corporate and higher education May webinar.pptx
Corporate and higher education May webinar.pptxCorporate and higher education May webinar.pptx
Corporate and higher education May webinar.pptx
 
Repurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost Saving
Repurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost SavingRepurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost Saving
Repurposing LNG terminals for Hydrogen Ammonia: Feasibility and Cost Saving
 
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
DEV meet-up UiPath Document Understanding May 7 2024 AmsterdamDEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
 
الأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهلهالأمن السيبراني - ما لا يسع للمستخدم جهله
الأمن السيبراني - ما لا يسع للمستخدم جهله
 
Quantum Leap in Next-Generation Computing
Quantum Leap in Next-Generation ComputingQuantum Leap in Next-Generation Computing
Quantum Leap in Next-Generation Computing
 
Six Myths about Ontologies: The Basics of Formal Ontology
Six Myths about Ontologies: The Basics of Formal OntologySix Myths about Ontologies: The Basics of Formal Ontology
Six Myths about Ontologies: The Basics of Formal Ontology
 
Modernizing Legacy Systems Using Ballerina
Modernizing Legacy Systems Using BallerinaModernizing Legacy Systems Using Ballerina
Modernizing Legacy Systems Using Ballerina
 
Simplifying Mobile A11y Presentation.pptx
Simplifying Mobile A11y Presentation.pptxSimplifying Mobile A11y Presentation.pptx
Simplifying Mobile A11y Presentation.pptx
 
Design and Development of a Provenance Capture Platform for Data Science
Design and Development of a Provenance Capture Platform for Data ScienceDesign and Development of a Provenance Capture Platform for Data Science
Design and Development of a Provenance Capture Platform for Data Science
 

Ensemble normalization for stable training

  • 1. Ensemble Normalization for Stable Training 𝑆𝑒𝑜𝑢𝑛𝑔-𝐻𝑜 𝐶ℎ𝑜𝑖 * and 𝐽𝑖 𝑊𝑜𝑜𝑛𝑔 𝐶ℎ𝑜𝑖 1. Department of Electronic Information Engineering, Hansung University 2. Department of Trade, Hansung University 1. Introduction Research on normalization is needed to correct and learn internal covariate shifts. Excessive internal covariate shifts can result in vanishing or exploding gradients. However, learning by conventional single normalization can be learned in an unoptimized direction. In order to guide the learning of different features in the right direction through learning, we propose an ensemble normalization that learns different features through two normalizations. Abstract Normalization studies have been conducted to solve vanishing or exploding gradient for stable training. This phenomenon leads to poor model training. Existing normalization method has used a single normalization. However, single normalization has limitations in stabilizing learning through the model weight of internal covariate shift prevention. For improvement, we propose new normalization method, an ensemble method that combines two normalization methods. We analyze the existing normalization method through an ablation study. To verify the proposed method, We experimented two models and two datasets in semantic segmentation task and two models and a dataset in generation task. As a result, we confirmed that the new normalization method is effective for improving stable training and semantic segmentation and generation performance. 2. Related Works Existing normalization studies have been conducted such as batch normalization (BN), instance normalization (IN), and group normalization (GN). BN normalizes samples in mini- batch units through using mean and variance of each mini-batch and also add scaling and shifting factors to the layers. However, BN does not perform well at small batch sizes. IN normalizes across each channel in each sample. IN performs well on style transfer regardless of the batch size. GN is similar to IN but normalizes over groups of channels for each sample. GN performed as well as BN on ImageNet. 3. Ensemble Normalization Figure 1. Process of Normalization Method, a) Existing normalization method, b) Ensemble normalization method Ensemble normalization calculates the addition of batch normalization and existing normalization. Then apply division 2 to the preceding result. The above calculation sequence is shown in Figure 1. Input is a group of feature maps convoluted by a convolutional layer. Output is a group of feature maps normalizaed by a normalization layer which is one of the conventional methods in a). However in b), the normalization layer is consist of two methods. To have a clear understanding of the existing methods, We conducted the ablation study on normalization to conduct an impact analysis on each existing normalization method before verifying the proposed method. In semantic segmentation task, we analyzed each normalization in the learning process through Focal loss, Hinge loss, and Cross entropy. Experimental results show average results from Focal loss, Hinge loss, and Cross entropy. The experimental method uses VOC and ATR datasets for FCN and U-Net and uses the average value of three losses and uses the average value through five iterative experiments. In generation task, we analyzed each normalization by using GAN and LSGAN models and MNIST dataset. 5. Conclusion We propose a new normalization which is an ensemble normalization. First, the impact of existing normalization was normalization was analyzed through ablation studies to determine the impact of the existing method. We also verified the ensemble normalization method proposed in this paper. As a result, we confirmed that the proposed method is effective for stable training on semantic segmentation and generation. a) b) FCN FCNIN FCNBN FCNGN Figure 2. Four feature maps in the second layer of the model, each with BN, IN, and GN applied. a) existing normalizations, b) the proposed normalization. As shown in figure 2, a comparison is made between the conventional normalization and the proposed method. In the case of existing normalization, each feature map is not different from each other, whereas in the feature map of the proposed method, different feature maps can be confirmed by looking at BIN (combined BN with IN) and BGN (combined BN with GN). This means that many different features have been trained, which leads to performance gains. In the case of BBN (combined BN with BN), on the other hand, three of the four feature maps show little difference. This means that rich features are not trained, and as shown as in table2, they do not perform relatively well. GAN GANBN GANGN GANBBN GANIN GANBIN GANBGN LSGAN LSGANBN LSGANGN LSGANBBN LSGANIN LSGANBIN LSGANBGN Figure 2. Four generated images of GAN and LSGAN with conventional normalization methods or ensemble normalization methods. 4. Experimental Results Table1. Quantitative comparison of two segmentation models(FCN and U-Net) using the VOC dataset and two generation models(GAN and LSGAN) using the MNIST dataset. Table 1 shows experimental results using two datasets for the existing methods and the proposed method and overall performance is improved. It is because they learn complementary to each other by catching the characteristics that cannot be captured by existing normalization methods through other normalization methods. With the addition of instance normalization, batch normalization, and group normalization, the loss each changes were +2.3%, +0.4%, and -4.2% for the FCN model and in the U-Net model, there were -42.9%, -20.3%, and -54.2%, respectively. Also, overall structure similarity(SSIM) values have been improved in the generator models. Through this experiment, we confirmed that the proposed method works well for the generator models. As shown as in figure 2, it is confirmed that By applying the proposed normalization method, the image is clearer than when applying the conventional normalization method.

Editor's Notes

  1. Figure 2에 나타나듯이 기존 노말라이제이션과 제안하는 방법을 비교해 보았을때, 두드러지는 특징이 발견 된다. 먼저 기존 노말라이제이션의 경우 각 feature map들이 서로 차이가 없는 반면에 제안하는 방법의 feature map들의 경우 BIN과 BGN을 보면 각기 다른 feature map들이 추출된 것을 확인 할 수 있다. 즉 서로 다른 여러가지의 특징이 훈련 되었다는 의미이며, 이는 곧 성능 향상으로 이어진다. 반면 BBN의 경우 네 장의 feature map중에서 세 장은 거의 차이가 없다. 이것은 풍부한 특징이 훈련되지 못한다는 의미이며, 이후 table2에서 확인할 수 있듯이 비교적 좋은 성능을 보이지 못한다.