SlideShare a Scribd company logo
1 of 20
Download to read offline
CANEXPOSURE,NOISEANDCOMPRESSION
AFFECTIMAGERECOGNITION?
An Assessment of the Impacts
on State-of-the-art ConvNets
STEFFENS, Cristiano R.; MESSIAS, Lucas R. V.;
DREWS-JR, Paulo J. L.;BOTELHO, Silvia S. d. C.
cristianosteffens@furg.br
OUTLINE
Evaluating image recognition models behiond validation sets
• Perception / Vision is an important component of modern autonomous systems
• CNNs hold the state-of-the-art in image recognition
• Growing interest in reliability / robustness
• Comprehensive assessment
• Clear methodology
• State-of-the-art models
• Several types of distortion
• Further directions
• How can we build better models?
• Can we prevent systems from operating on faulty data?
• Can we build better pipelines?
2
MOTIVATION
Under-exposure conditions
3
• Weakly illuminated scenes
• Time constraints (i.e. the robot depends on the image acquisition/processing to make a decision)
• Scenes with high dynamic ranges
• Small apperture (hardware construction)
• Low quality/cost sensors
Properly exposed Low Range Gamma 2 Gamma 4 Gamma 8
MOTIVATION
Some common conditions
4
MOTIVATION
Over-Exposure
5
• Scene with high dynamic range
• Ill adjusted optics/gain
• Time constraints
• Reflective surfaces
• Low dynamic range sensors
Properly exposed Low Range Gamma 1/2 Gamma 1/4 Gamma 1/8
MOTIVATION
Lossy Compression, Poisson, Gaussian, Salt & Pepper and Speckle Noise
6
• Bandwidth limitation
• Storage limitation
• Sensor quality
• Dead pixels (always off or always on)
• Wear and tear
• Dust, damage on lens and sensors, noise
Over-compression Poisson Noise Gaussian Noise Salt & Pepper Speckle Noise
EVALUATED MODELS
Imagenet Large Scale Visual Recognition Challenge (ILSVRC) State-of-the-Art holders
Model Year Size Parameters
Top-1
Acuracy
Input Size
Inception-
ResNet-v2
2017 215 MB 55,873,736 0,80 299
MobileNetV1 2017 16 MB 4,253,864 0,70 224
NASNetLarge 2018 343 MB 88,949,818 0,83 331
NASNetMobile 2018 23 MB 5,326,716 0,74 224
VGG16 2014 528 MB 138,357,544 0,71 224
Xception 2017 88 MB 22,910,480 0,79 299
7
PROCEDURE
A procedure that can be reproduced and used for any vision task
• We use pre-trained image recognition models
• No fine-tuning
• Exact same preprocessing as in the original implementation
• Oficial Imagenet validation set
• 1000 classes
• 50 images per class
• Inference on:
• Original set (to avoid hardware related, interpolation and other bias)
• 8 levels of misexposure
• Over-compressed images
• 4 types of typical noise
8
RESULTS
Overview – Top-1 Accuracy
9
RESULTS - INCEPTION-RESNET-V2
Overall good performance. Robust towards mild mis-exposure, compression, Gaussian and Poisson
10
RESULTS - INCEPTION-RESNET-V2
Overall good performance. Robust towards mild mis-exposure, compression, Gaussian and Poisson
11
FNs are limited to 50 due to the validation
dataset properties
No upper bound for FP
Statistics are per class:
A median of 10 means that 50% of the
classes in the dataset presented 10 or less
false negatives.
What is more important?
Would you rather overrun a person due to a
FN or stop in the middle of the road due to a
FP ?
RESULTS - MOBILENETV1
No robustness to S&P and Speckle Noise. Highly affected by moderate mis-exposure.
12
RESULTS - NASNET LARGE
Best accuracy, precision, and F1-Score among all models considered in this study
13
RESULTS - NASNET MOBILE
Significantly affected by severe miss-exposure conditions, S&P, and Speckle noise
14
RESULTS – VGG16
Slightly affected by mild mis-exposure and Poisson noise
15
RESULTS – XCEPTION
Robust towards moderate mis-exposure, over-compression, Gaussian and Poisson noise
16
CONCLUSION
New is Always better! Larger is better!
• Relevant
• Autonomous systems
• Robotics
• Applications that rely on visual perception
• Comprehensive experiment
• Broad set of classifiers
• Based on standard ILSVRC validation set
• Poor exposure
• Heavy compression
• Signal independent noise
• Signal dependent noise
• Reproducible procedure
• Objective evaluation
• No human bias
17
CONCLUSION
New is Always better! Larger is better!
• Most models are
• Little affected by mild miss-exposure.
• Robust towards Poisson and Gaussian noise
• Critically affected by moderate to severe miss exposure
• Critically affected by S&P and Speckle noise
• CNNs are evolving
• Modern architectures, such as NASNet, Inception Resnet v2 and Xception are more robust
• VGG is among the least robust
• Large models are better
• NOT you VGG!!
• NASNet Large performs significantly better than its Mobile version (while both share the same building
blocks)
• Mobile models are most affected
18
ONGOINGAND FUTURE WORK
We have a real issue! How can we solve it?
• Could the models’ accuracy be improved by adding these
common distortions in training time?
😞 Preliminary results show small improvement
• Can we build image processing pipelines which protect the
application from failing due to faulty data?
😃 Absolutelly! Preliminary results are promising 👉
• Can we prevent ill exposure in mobile/outdoor robotics?
⏳ Future Work
• Can we improve classification models by putting more
emphasis on image classes that are more prone to error?
⏳ Future Work
19
👈 Damaged
👈 Restored
☝️ Original
THANKS!
Cristiano Steffens
cristianosteffens@furg.br
researchgate.net/profile/Cristiano_Steffens
20
cristianosteffens@ieee.org.br
github.com/steffensbola
This study was financed in part by the Coordenação de Aperfeiçoamento de
Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001.

More Related Content

Similar to Can Exposure, Noise and Compression affect Image Recognition? An Assessment of the Impacts on State-of-the-art ConvNets

Face Detection.pptx
Face Detection.pptxFace Detection.pptx
Face Detection.pptxTorshaSett
 
Steer and/or sink the supertanker by Andrew Rendell
Steer and/or sink the supertanker by Andrew RendellSteer and/or sink the supertanker by Andrew Rendell
Steer and/or sink the supertanker by Andrew RendellValtech UK
 
Semi-Supervised Deep Learning
Semi-Supervised Deep LearningSemi-Supervised Deep Learning
Semi-Supervised Deep LearningKamer Ali Yuksel
 
Env2Vec: Accelerating VNF Testing with Deep Learning
Env2Vec: Accelerating VNF Testing with Deep LearningEnv2Vec: Accelerating VNF Testing with Deep Learning
Env2Vec: Accelerating VNF Testing with Deep LearningGUANGYUAN PIAO
 
Robustness of compressed CNNs
Robustness of compressed CNNsRobustness of compressed CNNs
Robustness of compressed CNNsKaushalya Madhawa
 
ROLE OF DIGITAL IMAGING IN PATHOLOGY.pptx
ROLE OF DIGITAL IMAGING IN PATHOLOGY.pptxROLE OF DIGITAL IMAGING IN PATHOLOGY.pptx
ROLE OF DIGITAL IMAGING IN PATHOLOGY.pptxaditisikarwar2
 
06 13sept 8313 9997-2-ed an adaptive (edit lafi)
06 13sept 8313 9997-2-ed an adaptive (edit lafi)06 13sept 8313 9997-2-ed an adaptive (edit lafi)
06 13sept 8313 9997-2-ed an adaptive (edit lafi)IAESIJEECS
 
Convolutional neural networks 이론과 응용
Convolutional neural networks 이론과 응용Convolutional neural networks 이론과 응용
Convolutional neural networks 이론과 응용홍배 김
 
“A Highly Data-Efficient Deep Learning Approach,” a Presentation from Samsung
“A Highly Data-Efficient Deep Learning Approach,” a Presentation from Samsung“A Highly Data-Efficient Deep Learning Approach,” a Presentation from Samsung
“A Highly Data-Efficient Deep Learning Approach,” a Presentation from SamsungEdge AI and Vision Alliance
 
SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...
SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...
SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...NAVER Engineering
 
SCREENLESS DISPLAY.pptx
SCREENLESS DISPLAY.pptxSCREENLESS DISPLAY.pptx
SCREENLESS DISPLAY.pptxAlenJames14
 
A DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGES
A DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGESA DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGES
A DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGESPNandaSai
 
DATI, AI E ROBOTICA @POLITO
DATI, AI E ROBOTICA @POLITODATI, AI E ROBOTICA @POLITO
DATI, AI E ROBOTICA @POLITOMarcoMellia
 
IRJET- Exploring Image Super Resolution Techniques
IRJET- Exploring Image Super Resolution TechniquesIRJET- Exploring Image Super Resolution Techniques
IRJET- Exploring Image Super Resolution TechniquesIRJET Journal
 
Fcv core liu
Fcv core liuFcv core liu
Fcv core liuzukun
 
Jillian ms defense-4-14-14-ja
Jillian ms defense-4-14-14-jaJillian ms defense-4-14-14-ja
Jillian ms defense-4-14-14-jaJillian Aurisano
 
Virtual Retinal Display: their falling cost and rising performance
Virtual Retinal Display: their falling cost and rising performanceVirtual Retinal Display: their falling cost and rising performance
Virtual Retinal Display: their falling cost and rising performanceJeffrey Funk
 
AI Class Topic 6: Easy Way to Learn Deep Learning AI Technologies
AI Class Topic 6: Easy Way to Learn Deep Learning AI TechnologiesAI Class Topic 6: Easy Way to Learn Deep Learning AI Technologies
AI Class Topic 6: Easy Way to Learn Deep Learning AI TechnologiesValue Amplify Consulting
 

Similar to Can Exposure, Noise and Compression affect Image Recognition? An Assessment of the Impacts on State-of-the-art ConvNets (20)

Face Detection.pptx
Face Detection.pptxFace Detection.pptx
Face Detection.pptx
 
Steer and/or sink the supertanker by Andrew Rendell
Steer and/or sink the supertanker by Andrew RendellSteer and/or sink the supertanker by Andrew Rendell
Steer and/or sink the supertanker by Andrew Rendell
 
Semi-Supervised Deep Learning
Semi-Supervised Deep LearningSemi-Supervised Deep Learning
Semi-Supervised Deep Learning
 
Env2Vec: Accelerating VNF Testing with Deep Learning
Env2Vec: Accelerating VNF Testing with Deep LearningEnv2Vec: Accelerating VNF Testing with Deep Learning
Env2Vec: Accelerating VNF Testing with Deep Learning
 
Robustness of compressed CNNs
Robustness of compressed CNNsRobustness of compressed CNNs
Robustness of compressed CNNs
 
ROLE OF DIGITAL IMAGING IN PATHOLOGY.pptx
ROLE OF DIGITAL IMAGING IN PATHOLOGY.pptxROLE OF DIGITAL IMAGING IN PATHOLOGY.pptx
ROLE OF DIGITAL IMAGING IN PATHOLOGY.pptx
 
06 13sept 8313 9997-2-ed an adaptive (edit lafi)
06 13sept 8313 9997-2-ed an adaptive (edit lafi)06 13sept 8313 9997-2-ed an adaptive (edit lafi)
06 13sept 8313 9997-2-ed an adaptive (edit lafi)
 
Convolutional neural networks 이론과 응용
Convolutional neural networks 이론과 응용Convolutional neural networks 이론과 응용
Convolutional neural networks 이론과 응용
 
“A Highly Data-Efficient Deep Learning Approach,” a Presentation from Samsung
“A Highly Data-Efficient Deep Learning Approach,” a Presentation from Samsung“A Highly Data-Efficient Deep Learning Approach,” a Presentation from Samsung
“A Highly Data-Efficient Deep Learning Approach,” a Presentation from Samsung
 
SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...
SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...
SafeguardAI and Surprise Based Learning -- Protect your AI solutions from Uni...
 
Senslab - open hardware - fossa2010
Senslab - open hardware - fossa2010Senslab - open hardware - fossa2010
Senslab - open hardware - fossa2010
 
SCREENLESS DISPLAY.pptx
SCREENLESS DISPLAY.pptxSCREENLESS DISPLAY.pptx
SCREENLESS DISPLAY.pptx
 
A DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGES
A DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGESA DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGES
A DEEP LEARNING APPROACH FOR SEMANTIC SEGMENTATION IN BRAIN TUMOR IMAGES
 
DATI, AI E ROBOTICA @POLITO
DATI, AI E ROBOTICA @POLITODATI, AI E ROBOTICA @POLITO
DATI, AI E ROBOTICA @POLITO
 
IRJET- Exploring Image Super Resolution Techniques
IRJET- Exploring Image Super Resolution TechniquesIRJET- Exploring Image Super Resolution Techniques
IRJET- Exploring Image Super Resolution Techniques
 
Fcv core liu
Fcv core liuFcv core liu
Fcv core liu
 
Jillian ms defense-4-14-14-ja
Jillian ms defense-4-14-14-jaJillian ms defense-4-14-14-ja
Jillian ms defense-4-14-14-ja
 
Virtual Retinal Display: their falling cost and rising performance
Virtual Retinal Display: their falling cost and rising performanceVirtual Retinal Display: their falling cost and rising performance
Virtual Retinal Display: their falling cost and rising performance
 
AI Class Topic 6: Easy Way to Learn Deep Learning AI Technologies
AI Class Topic 6: Easy Way to Learn Deep Learning AI TechnologiesAI Class Topic 6: Easy Way to Learn Deep Learning AI Technologies
AI Class Topic 6: Easy Way to Learn Deep Learning AI Technologies
 
Seminar nov2017
Seminar nov2017Seminar nov2017
Seminar nov2017
 

More from Cristiano Rafael Steffens

CONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and video
CONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and videoCONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and video
CONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and videoCristiano Rafael Steffens
 
A CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGES
A CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGESA CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGES
A CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGESCristiano Rafael Steffens
 
MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...
MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...
MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...Cristiano Rafael Steffens
 
UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...
UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...
UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...Cristiano Rafael Steffens
 
FPGA-based sensor integration and communication protocols for automated
FPGA-based sensor integration and communication protocols for automatedFPGA-based sensor integration and communication protocols for automated
FPGA-based sensor integration and communication protocols for automatedCristiano Rafael Steffens
 
Lars 2016 A Texture Driven Approach for Visible Spectrum Fire Detection
Lars 2016 A Texture Driven Approach for Visible Spectrum Fire DetectionLars 2016 A Texture Driven Approach for Visible Spectrum Fire Detection
Lars 2016 A Texture Driven Approach for Visible Spectrum Fire DetectionCristiano Rafael Steffens
 
ICRA 2016 - Interactive section Presentation
ICRA 2016 - Interactive section PresentationICRA 2016 - Interactive section Presentation
ICRA 2016 - Interactive section PresentationCristiano Rafael Steffens
 
Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...
Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...
Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...Cristiano Rafael Steffens
 
Simpósio Unicruz: OpenCV + Python (parte 1)
Simpósio Unicruz: OpenCV + Python (parte 1)Simpósio Unicruz: OpenCV + Python (parte 1)
Simpósio Unicruz: OpenCV + Python (parte 1)Cristiano Rafael Steffens
 
Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...
Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...
Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...Cristiano Rafael Steffens
 
Automated control module based on VBM for shipyard welding applications: Stud...
Automated control module based on VBM for shipyard welding applications: Stud...Automated control module based on VBM for shipyard welding applications: Stud...
Automated control module based on VBM for shipyard welding applications: Stud...Cristiano Rafael Steffens
 
An Unconstrained Dataset for Non-stationary Video Based Fire Detection
An Unconstrained Dataset for Non-stationary Video Based Fire DetectionAn Unconstrained Dataset for Non-stationary Video Based Fire Detection
An Unconstrained Dataset for Non-stationary Video Based Fire DetectionCristiano Rafael Steffens
 
Introdução ao processamento de imagens com OpenCV (cont)
Introdução ao processamento de imagens com OpenCV (cont)Introdução ao processamento de imagens com OpenCV (cont)
Introdução ao processamento de imagens com OpenCV (cont)Cristiano Rafael Steffens
 
Um Sistema De Detecção De Fogo Baseado Em Vídeo
Um Sistema De Detecção De Fogo Baseado Em VídeoUm Sistema De Detecção De Fogo Baseado Em Vídeo
Um Sistema De Detecção De Fogo Baseado Em VídeoCristiano Rafael Steffens
 
Um sistema de detecção de chamas utilizando RF e SVM (Short Version)
Um sistema de detecção de chamas utilizando RF e SVM (Short Version)Um sistema de detecção de chamas utilizando RF e SVM (Short Version)
Um sistema de detecção de chamas utilizando RF e SVM (Short Version)Cristiano Rafael Steffens
 

More from Cristiano Rafael Steffens (20)

CONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and video
CONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and videoCONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and video
CONVOLUTIONAL NEURAL NETWORKS: The workhorse of image and video
 
A CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGES
A CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGESA CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGES
A CNN BASED MODEL TO RESTORE ILL EXPOSED IMAGES
 
MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...
MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...
MODELAGEM DAS DINÂMICAS DA FORMAÇÃO DA GOTA E TRANSFERÊNCIA DE MASSA EM PROCE...
 
UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...
UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...
UMA ABORDAGEM COMPARATIVA ENTRE MICROCONTROLADORES: ARDUINO MEGA X ARDUINO DU...
 
FPGA-based sensor integration and communication protocols for automated
FPGA-based sensor integration and communication protocols for automatedFPGA-based sensor integration and communication protocols for automated
FPGA-based sensor integration and communication protocols for automated
 
Lars 2016 A Texture Driven Approach for Visible Spectrum Fire Detection
Lars 2016 A Texture Driven Approach for Visible Spectrum Fire DetectionLars 2016 A Texture Driven Approach for Visible Spectrum Fire Detection
Lars 2016 A Texture Driven Approach for Visible Spectrum Fire Detection
 
Php Math and arrays
Php Math and arraysPhp Math and arrays
Php Math and arrays
 
ICRA 2016 - Interactive section Presentation
ICRA 2016 - Interactive section PresentationICRA 2016 - Interactive section Presentation
ICRA 2016 - Interactive section Presentation
 
Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...
Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...
Vision-Based System for Welding Groove Measurements for Robotic Welding Appli...
 
Simpósio Unicruz: OpenCV + Python (parte 1)
Simpósio Unicruz: OpenCV + Python (parte 1)Simpósio Unicruz: OpenCV + Python (parte 1)
Simpósio Unicruz: OpenCV + Python (parte 1)
 
Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...
Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...
Welding Groove Mapping: Image Acquisition and Processing on Shiny Surfaces - ...
 
Automated control module based on VBM for shipyard welding applications: Stud...
Automated control module based on VBM for shipyard welding applications: Stud...Automated control module based on VBM for shipyard welding applications: Stud...
Automated control module based on VBM for shipyard welding applications: Stud...
 
An Unconstrained Dataset for Non-stationary Video Based Fire Detection
An Unconstrained Dataset for Non-stationary Video Based Fire DetectionAn Unconstrained Dataset for Non-stationary Video Based Fire Detection
An Unconstrained Dataset for Non-stationary Video Based Fire Detection
 
Introdução ao processamento de imagens com OpenCV (cont)
Introdução ao processamento de imagens com OpenCV (cont)Introdução ao processamento de imagens com OpenCV (cont)
Introdução ao processamento de imagens com OpenCV (cont)
 
Introdução OpenCV (Pt-Br) com exemplos
Introdução OpenCV (Pt-Br) com exemplosIntrodução OpenCV (Pt-Br) com exemplos
Introdução OpenCV (Pt-Br) com exemplos
 
Um Sistema De Detecção De Fogo Baseado Em Vídeo
Um Sistema De Detecção De Fogo Baseado Em VídeoUm Sistema De Detecção De Fogo Baseado Em Vídeo
Um Sistema De Detecção De Fogo Baseado Em Vídeo
 
Um sistema de detecção de chamas utilizando RF e SVM (Short Version)
Um sistema de detecção de chamas utilizando RF e SVM (Short Version)Um sistema de detecção de chamas utilizando RF e SVM (Short Version)
Um sistema de detecção de chamas utilizando RF e SVM (Short Version)
 
G xserver curso-actualizgxxev1
G xserver curso-actualizgxxev1G xserver curso-actualizgxxev1
G xserver curso-actualizgxxev1
 
20 pratico cursogxxbr
20 pratico cursogxxbr20 pratico cursogxxbr
20 pratico cursogxxbr
 
18 indice cursogxx
18 indice cursogxx18 indice cursogxx
18 indice cursogxx
 

Recently uploaded

POGONATUM : morphology, anatomy, reproduction etc.
POGONATUM : morphology, anatomy, reproduction etc.POGONATUM : morphology, anatomy, reproduction etc.
POGONATUM : morphology, anatomy, reproduction etc.Cherry
 
Energy is the beat of life irrespective of the domains. ATP- the energy curre...
Energy is the beat of life irrespective of the domains. ATP- the energy curre...Energy is the beat of life irrespective of the domains. ATP- the energy curre...
Energy is the beat of life irrespective of the domains. ATP- the energy curre...Nistarini College, Purulia (W.B) India
 
FAIRSpectra - Enabling the FAIRification of Analytical Science
FAIRSpectra - Enabling the FAIRification of Analytical ScienceFAIRSpectra - Enabling the FAIRification of Analytical Science
FAIRSpectra - Enabling the FAIRification of Analytical ScienceAlex Henderson
 
Dr. E. Muralinath_ Blood indices_clinical aspects
Dr. E. Muralinath_ Blood indices_clinical  aspectsDr. E. Muralinath_ Blood indices_clinical  aspects
Dr. E. Muralinath_ Blood indices_clinical aspectsmuralinath2
 
Terpineol and it's characterization pptx
Terpineol and it's characterization pptxTerpineol and it's characterization pptx
Terpineol and it's characterization pptxMuhammadRazzaq31
 
GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry
GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry
GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry Areesha Ahmad
 
Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....
Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....
Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....muralinath2
 
Cyathodium bryophyte: morphology, anatomy, reproduction etc.
Cyathodium bryophyte: morphology, anatomy, reproduction etc.Cyathodium bryophyte: morphology, anatomy, reproduction etc.
Cyathodium bryophyte: morphology, anatomy, reproduction etc.Cherry
 
module for grade 9 for distance learning
module for grade 9 for distance learningmodule for grade 9 for distance learning
module for grade 9 for distance learninglevieagacer
 
ONLINE VOTING SYSTEM SE Project for vote
ONLINE VOTING SYSTEM SE Project for voteONLINE VOTING SYSTEM SE Project for vote
ONLINE VOTING SYSTEM SE Project for voteRaunakRastogi4
 
Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.
Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.
Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.Cherry
 
Digital Dentistry.Digital Dentistryvv.pptx
Digital Dentistry.Digital Dentistryvv.pptxDigital Dentistry.Digital Dentistryvv.pptx
Digital Dentistry.Digital Dentistryvv.pptxMohamedFarag457087
 
Plasmid: types, structure and functions.
Plasmid: types, structure and functions.Plasmid: types, structure and functions.
Plasmid: types, structure and functions.Cherry
 
FAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and SpectrometryFAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and SpectrometryAlex Henderson
 
Pteris : features, anatomy, morphology and lifecycle
Pteris : features, anatomy, morphology and lifecyclePteris : features, anatomy, morphology and lifecycle
Pteris : features, anatomy, morphology and lifecycleCherry
 
Porella : features, morphology, anatomy, reproduction etc.
Porella : features, morphology, anatomy, reproduction etc.Porella : features, morphology, anatomy, reproduction etc.
Porella : features, morphology, anatomy, reproduction etc.Cherry
 
CYTOGENETIC MAP................ ppt.pptx
CYTOGENETIC MAP................ ppt.pptxCYTOGENETIC MAP................ ppt.pptx
CYTOGENETIC MAP................ ppt.pptxCherry
 
Cot curve, melting temperature, unique and repetitive DNA
Cot curve, melting temperature, unique and repetitive DNACot curve, melting temperature, unique and repetitive DNA
Cot curve, melting temperature, unique and repetitive DNACherry
 
Genome sequencing,shotgun sequencing.pptx
Genome sequencing,shotgun sequencing.pptxGenome sequencing,shotgun sequencing.pptx
Genome sequencing,shotgun sequencing.pptxCherry
 
Efficient spin-up of Earth System Models usingsequence acceleration
Efficient spin-up of Earth System Models usingsequence accelerationEfficient spin-up of Earth System Models usingsequence acceleration
Efficient spin-up of Earth System Models usingsequence accelerationSérgio Sacani
 

Recently uploaded (20)

POGONATUM : morphology, anatomy, reproduction etc.
POGONATUM : morphology, anatomy, reproduction etc.POGONATUM : morphology, anatomy, reproduction etc.
POGONATUM : morphology, anatomy, reproduction etc.
 
Energy is the beat of life irrespective of the domains. ATP- the energy curre...
Energy is the beat of life irrespective of the domains. ATP- the energy curre...Energy is the beat of life irrespective of the domains. ATP- the energy curre...
Energy is the beat of life irrespective of the domains. ATP- the energy curre...
 
FAIRSpectra - Enabling the FAIRification of Analytical Science
FAIRSpectra - Enabling the FAIRification of Analytical ScienceFAIRSpectra - Enabling the FAIRification of Analytical Science
FAIRSpectra - Enabling the FAIRification of Analytical Science
 
Dr. E. Muralinath_ Blood indices_clinical aspects
Dr. E. Muralinath_ Blood indices_clinical  aspectsDr. E. Muralinath_ Blood indices_clinical  aspects
Dr. E. Muralinath_ Blood indices_clinical aspects
 
Terpineol and it's characterization pptx
Terpineol and it's characterization pptxTerpineol and it's characterization pptx
Terpineol and it's characterization pptx
 
GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry
GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry
GBSN - Biochemistry (Unit 2) Basic concept of organic chemistry
 
Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....
Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....
Human & Veterinary Respiratory Physilogy_DR.E.Muralinath_Associate Professor....
 
Cyathodium bryophyte: morphology, anatomy, reproduction etc.
Cyathodium bryophyte: morphology, anatomy, reproduction etc.Cyathodium bryophyte: morphology, anatomy, reproduction etc.
Cyathodium bryophyte: morphology, anatomy, reproduction etc.
 
module for grade 9 for distance learning
module for grade 9 for distance learningmodule for grade 9 for distance learning
module for grade 9 for distance learning
 
ONLINE VOTING SYSTEM SE Project for vote
ONLINE VOTING SYSTEM SE Project for voteONLINE VOTING SYSTEM SE Project for vote
ONLINE VOTING SYSTEM SE Project for vote
 
Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.
Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.
Genome Projects : Human, Rice,Wheat,E coli and Arabidopsis.
 
Digital Dentistry.Digital Dentistryvv.pptx
Digital Dentistry.Digital Dentistryvv.pptxDigital Dentistry.Digital Dentistryvv.pptx
Digital Dentistry.Digital Dentistryvv.pptx
 
Plasmid: types, structure and functions.
Plasmid: types, structure and functions.Plasmid: types, structure and functions.
Plasmid: types, structure and functions.
 
FAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and SpectrometryFAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
FAIRSpectra - Enabling the FAIRification of Spectroscopy and Spectrometry
 
Pteris : features, anatomy, morphology and lifecycle
Pteris : features, anatomy, morphology and lifecyclePteris : features, anatomy, morphology and lifecycle
Pteris : features, anatomy, morphology and lifecycle
 
Porella : features, morphology, anatomy, reproduction etc.
Porella : features, morphology, anatomy, reproduction etc.Porella : features, morphology, anatomy, reproduction etc.
Porella : features, morphology, anatomy, reproduction etc.
 
CYTOGENETIC MAP................ ppt.pptx
CYTOGENETIC MAP................ ppt.pptxCYTOGENETIC MAP................ ppt.pptx
CYTOGENETIC MAP................ ppt.pptx
 
Cot curve, melting temperature, unique and repetitive DNA
Cot curve, melting temperature, unique and repetitive DNACot curve, melting temperature, unique and repetitive DNA
Cot curve, melting temperature, unique and repetitive DNA
 
Genome sequencing,shotgun sequencing.pptx
Genome sequencing,shotgun sequencing.pptxGenome sequencing,shotgun sequencing.pptx
Genome sequencing,shotgun sequencing.pptx
 
Efficient spin-up of Earth System Models usingsequence acceleration
Efficient spin-up of Earth System Models usingsequence accelerationEfficient spin-up of Earth System Models usingsequence acceleration
Efficient spin-up of Earth System Models usingsequence acceleration
 

Can Exposure, Noise and Compression affect Image Recognition? An Assessment of the Impacts on State-of-the-art ConvNets

  • 1. CANEXPOSURE,NOISEANDCOMPRESSION AFFECTIMAGERECOGNITION? An Assessment of the Impacts on State-of-the-art ConvNets STEFFENS, Cristiano R.; MESSIAS, Lucas R. V.; DREWS-JR, Paulo J. L.;BOTELHO, Silvia S. d. C. cristianosteffens@furg.br
  • 2. OUTLINE Evaluating image recognition models behiond validation sets • Perception / Vision is an important component of modern autonomous systems • CNNs hold the state-of-the-art in image recognition • Growing interest in reliability / robustness • Comprehensive assessment • Clear methodology • State-of-the-art models • Several types of distortion • Further directions • How can we build better models? • Can we prevent systems from operating on faulty data? • Can we build better pipelines? 2
  • 3. MOTIVATION Under-exposure conditions 3 • Weakly illuminated scenes • Time constraints (i.e. the robot depends on the image acquisition/processing to make a decision) • Scenes with high dynamic ranges • Small apperture (hardware construction) • Low quality/cost sensors Properly exposed Low Range Gamma 2 Gamma 4 Gamma 8
  • 5. MOTIVATION Over-Exposure 5 • Scene with high dynamic range • Ill adjusted optics/gain • Time constraints • Reflective surfaces • Low dynamic range sensors Properly exposed Low Range Gamma 1/2 Gamma 1/4 Gamma 1/8
  • 6. MOTIVATION Lossy Compression, Poisson, Gaussian, Salt & Pepper and Speckle Noise 6 • Bandwidth limitation • Storage limitation • Sensor quality • Dead pixels (always off or always on) • Wear and tear • Dust, damage on lens and sensors, noise Over-compression Poisson Noise Gaussian Noise Salt & Pepper Speckle Noise
  • 7. EVALUATED MODELS Imagenet Large Scale Visual Recognition Challenge (ILSVRC) State-of-the-Art holders Model Year Size Parameters Top-1 Acuracy Input Size Inception- ResNet-v2 2017 215 MB 55,873,736 0,80 299 MobileNetV1 2017 16 MB 4,253,864 0,70 224 NASNetLarge 2018 343 MB 88,949,818 0,83 331 NASNetMobile 2018 23 MB 5,326,716 0,74 224 VGG16 2014 528 MB 138,357,544 0,71 224 Xception 2017 88 MB 22,910,480 0,79 299 7
  • 8. PROCEDURE A procedure that can be reproduced and used for any vision task • We use pre-trained image recognition models • No fine-tuning • Exact same preprocessing as in the original implementation • Oficial Imagenet validation set • 1000 classes • 50 images per class • Inference on: • Original set (to avoid hardware related, interpolation and other bias) • 8 levels of misexposure • Over-compressed images • 4 types of typical noise 8
  • 10. RESULTS - INCEPTION-RESNET-V2 Overall good performance. Robust towards mild mis-exposure, compression, Gaussian and Poisson 10
  • 11. RESULTS - INCEPTION-RESNET-V2 Overall good performance. Robust towards mild mis-exposure, compression, Gaussian and Poisson 11 FNs are limited to 50 due to the validation dataset properties No upper bound for FP Statistics are per class: A median of 10 means that 50% of the classes in the dataset presented 10 or less false negatives. What is more important? Would you rather overrun a person due to a FN or stop in the middle of the road due to a FP ?
  • 12. RESULTS - MOBILENETV1 No robustness to S&P and Speckle Noise. Highly affected by moderate mis-exposure. 12
  • 13. RESULTS - NASNET LARGE Best accuracy, precision, and F1-Score among all models considered in this study 13
  • 14. RESULTS - NASNET MOBILE Significantly affected by severe miss-exposure conditions, S&P, and Speckle noise 14
  • 15. RESULTS – VGG16 Slightly affected by mild mis-exposure and Poisson noise 15
  • 16. RESULTS – XCEPTION Robust towards moderate mis-exposure, over-compression, Gaussian and Poisson noise 16
  • 17. CONCLUSION New is Always better! Larger is better! • Relevant • Autonomous systems • Robotics • Applications that rely on visual perception • Comprehensive experiment • Broad set of classifiers • Based on standard ILSVRC validation set • Poor exposure • Heavy compression • Signal independent noise • Signal dependent noise • Reproducible procedure • Objective evaluation • No human bias 17
  • 18. CONCLUSION New is Always better! Larger is better! • Most models are • Little affected by mild miss-exposure. • Robust towards Poisson and Gaussian noise • Critically affected by moderate to severe miss exposure • Critically affected by S&P and Speckle noise • CNNs are evolving • Modern architectures, such as NASNet, Inception Resnet v2 and Xception are more robust • VGG is among the least robust • Large models are better • NOT you VGG!! • NASNet Large performs significantly better than its Mobile version (while both share the same building blocks) • Mobile models are most affected 18
  • 19. ONGOINGAND FUTURE WORK We have a real issue! How can we solve it? • Could the models’ accuracy be improved by adding these common distortions in training time? 😞 Preliminary results show small improvement • Can we build image processing pipelines which protect the application from failing due to faulty data? 😃 Absolutelly! Preliminary results are promising 👉 • Can we prevent ill exposure in mobile/outdoor robotics? ⏳ Future Work • Can we improve classification models by putting more emphasis on image classes that are more prone to error? ⏳ Future Work 19 👈 Damaged 👈 Restored ☝️ Original
  • 20. THANKS! Cristiano Steffens cristianosteffens@furg.br researchgate.net/profile/Cristiano_Steffens 20 cristianosteffens@ieee.org.br github.com/steffensbola This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001.