SlideShare ist ein Scribd-Unternehmen logo
1 von 31
Downloaden Sie, um offline zu lesen
Democratizing Object Detection
BigML Release: Object Detection
BigML, Inc 2
Object Detection
Enter questions into chat box. We will answer some via chat and others
at the end of the session.
QUESTIONS
info@bigml.com @bigmlcom
Atakan Cetinsoy


VP of Predictive Applications.
MODERATOR
Charles Parker Ph.D.


VP of Machine Learning Algorithms.
SPEAKER
https://bigml.com/releases/object-detection
RESOURCES
BigML Release: Object Detection
BigML, Inc 3
Object Detection is a fundamental Computer Vision task that
involves localizing and classifying objects-of-interest in an image.
Introducing Object Detection
BigML Release: Object Detection
BigML, Inc 4
Object Detection
truck
car
Built on the foundation of BigML Image Processing, Object Detection allows you to localize
and classify objects in a fast and scalable manner.
Image Classi
fi
cation vs. Object Detection
Image Classi
fi
cation
BigML Release: Object Detection
BigML, Inc 5
Maturity of DL
techniques
Decreasing
compute costs
Abundance 

of image data
High-speed 

processing
(frames/sec.)
Next-Gen Tools
Object Detection - Why now?
BigML Release: Object Detection
BigML, Inc 6
Object Detection is Everywhere!
Healthcare
Agriculture
Manufacturing
Sports Transportation
Retail
BigML Release: Object Detection
BigML, Inc 7
Get plate numbers and states from distinct highway toll car images
Toll Video Cameras
1
Individual Car Images
2
Plate Detection
3
BigML Object Detection
Plate Extraction
4
5 Characters Recognition
6 State Classi
fi
er
BigML Object Detection
Character Images Classi
fi
cation
Image Classi
fi
cation
7
Automatic
Payment
After detecting plate
number and state,
automatic payment
is possible
Object Detection in Transportation
BigML Release: Object Detection
BigML, Inc 8
Smart Video Review Framework: Find, Group, Label, and Review
Find interesting events
Flags
Chants
Group events
Experts review clusters 

and assign their own labels
1 2
Find anomalous events
Cluster similar events
Smoke bombs / Flares
Label event groups
3
4
Assign labels to
known events
Models learn new
labels from expert’s
feedback
Object Detection for Public Safety
BigML Release: Object Detection
BigML, Inc 9
Object Detection without BigML
Image Data
Labeled Image Data
Labeling Tool Modeling Tool
Predictions
Con
fi
gured


Machine w/ GPU
Modeling Hardware


or Service
Trained Model
LABELERS
ML ENGINEER
IT DEPARTMENT
BigML Release: Object Detection
BigML, Inc 10
Scaling Object Detection without BigML
Image Data
Labeled Image Data
Trained Model
Con
fi
gure
d

Machine w/ GPU
Labeling Tool Modeling Tool
Modeling
Hardware or
Predictions
Con
fi
gure
d

Machine w/ GPU
Con
fi
gure
d

Machine w/ GPU
Con
fi
gure
d

Machine w/ GPU
Con
fi
gure
d

Machine w/ GPU
Con
fi
gured


Machine w/ GPU
Modeling
Hardware or
Modeling
Hardware or
Modeling
Hardware or
Modeling Hardware


or Service
MLOps / Reporting
Trained Model
Trained Model
LABELERS
ML ENGINEER
IT DEPARTMENT
BigML Release: Object Detection
BigML, Inc 11
Scalable Object Detection with BigML
LABELERS
ML ENGINEER
Image Data
Predictions
BigML Release: Object Detection
BigML, Inc 12
BigML Composability
ML Work
fl
ow
Object Detection
BigML Resources High-Level Algorithm 1-Click Execution
BigML Release: Object Detection
BigML, Inc 13
BigML Resource Synergies
BigML Release: Object Detection
BigML, Inc 14
Handling Images the BigML Way
• ALL the tasks for generating insights from
image data ON A SINGLE PLATFORM:
From labeling to inference, evaluation, and
predictions. 


• Streamlined image dataset management
with composite sources.


• Automatic handling of infrastructure
concerns.


• Your choice: Code and no code.
Solving image data-driven business problems with remarkable ease of use
BigML Release: Object Detection
BigML, Inc 15
What Object Detection Is
• Simultaneously locating and classifying specific sub-areas
(returned as xmin, ymin, xmax, ymax) in a given image.


• Many uses, but especially in:


• Manufacturing defect detection


• Medical image analysis


• Autonomous navigation


• Security (monitoring) applications
BigML Release: Object Detection
BigML, Inc 16
What Object Detection Is
• Identification of a specific area in the image that contains
an object and identification of the class of the object.


• The “score” given to the box is a product of the probability
of both of these factors. Note that even if the model
thinks there’s an object in a particular region, it can still fail
if the model has trouble deducing the class of the object.


• While we typically visualize the prediction as a box, this is
just for convenience.


• The parameters xmin, ymin, xmax, ymax could also specify an ellipse


• The boundaries of the object are unlikely to be pixel-accurate, even for
objects that are axis-aligned rectangles


• Better to think of the returned values is “approximate location” and
“approximate size”
BigML Release: Object Detection
BigML, Inc 17
What Object Detection Isn’t
• Object detection tells you the class for a specific
window of the image. If you want to classify the whole
image, you want “image classification” (available on
BigML).


• Object detection does not give an idea of exactly
which pixels in the image are in the detected object.
This is done using one of the various flavors of “image
segmentation” algorithms.


• Object detection on its own cannot find specific parts
of the object being detected. This is “keypoint
detection”, or for articulated objects (like people),
“pose estimation”. Though it is possible to label and
detect parts separately with object detection.
BigML Release: Object Detection
BigML, Inc 18
New Optype: Regions
Type Description Example
Numeric A continuous numeric value 0,75
Categorical One of a small set of possible values Cat
Datetime A String specifying a date, time, or both 4-22-2022 11:24am
Text An unstructured string of free text “Hello, cats!”
Image A
fi
le containing an image
Path The path of the image
fi
le in the zip
fi
le containing it cat/image1.jpg
Regions A list of
fi
ve-tuples, [label, xmin, ymin, xmax, ymax]
[[“cat”, 0.5, 0.1, 0.6, 0.9],
 

[“dog”, 0.2, 0.2, 0.8, 0.8]]
BigML, Inc BigML Release: Object Detection
Regions Labeling
19
BigML Release: Object Detection
BigML, Inc 20
The More Things Change . . .
Training Dataset
Test Dataset
Zip
fi
le with Images
Train
Evaluate
BigML, Inc BigML Release: Object Detection
Training An Object Detector
21
BigML Release: Object Detection
BigML, Inc 22
Under The Hood
• Break the image into a rough grid of cells.


• Centered at each cell, consider a number of
“bounding box proposals”.


• For each box in each cell, try to predict the
following things simultaneously:


• Whether or not there’s a object inside.


• The class of the object.


• Adjustments needed for the height/width/x-center/y-
center to make the box fit properly around the object.


• The shapes of the proposals are learned
directly from the data.
BigML Release: Object Detection
BigML, Inc 23
Some Caveats
• The usual caveats about feeding images to
machine learning models still apply.


• Example: Blurring.


• Example: Negative images.


• The bounding box proposals are learned
from the sizes given in your data.


• The boxes consider the entire image when
making predictions, not just the content of
the box proposal itself.
BigML Release: Object Detection
BigML, Inc 24
What if we Need More Power?
• For the public deployment, we use a small object detection architecture which performs reasonably well on most
problems.


• There are, however, other architectures that can give a significant bump in accuracy.


• They need significantly more data to train properly.


• They are slower for both training and testing.


• We don’t make these available on our public deployment, but contact us about these options in private
deployments.
BigML, Inc BigML Release: Object Detection
Making Predictions with Your Detector
25
BigML Release: Object Detection
BigML, Inc 26
Evaluations - A Little Different
• In classification problems, we can construct a “confusion matrix”
with true positives, true negatives, false positives, and false
negatives.


• For object detection, we don’t typically count “true negatives” as we
don’t give the model credit for predicting “nothing” for the vast
majority of locations, which would dominate metrics that consider
them.


• So we must rely on metrics that don’t involve true negatives, which
are based on precision, recall, and their combinations (e.g., average
precision; the precision averaged over several levels of recall).


• We can still change the score threshold for the model, and see how
that impacts the count of true positives, false positives, and false
negatives.
BigML Release: Object Detection
BigML, Inc 27
What is a “Correct” Prediction?
• Clearly, if the boxes align exactly, the prediction is correct, but they usually don’t.


• For each possible match, we compute a quantity called “Intersection over Union” or IOU.


• If the IOU is greater than some threshold, we consider the prediction correct.


• We do this for several thresholds in the evaluation.
BigML, Inc BigML Release: Object Detection
Object Detection Evaluation
28
BigML Release: Object Detection
BigML, Inc
Object Detection Model
Classi
fi
er
29
A Simple Work
fl
ow
Anomaly Detector
Object 1 Object 2
Object Classes Anomaly Scores
• What if you have a basic list of things you would
like to detect, but you know it’s not exhaustive?


• This might cause low scores for rare / never
seen objects.


• Solution: Train an object detector to detect a
generic “thing”.


• Crop out the detected objects, then passed the
crops to both an anomaly detector and a
classifier.


• If it’s a known object, you’ll get a high-
confidence class. If it’s a strange object, you will
get a high anomaly score.
BigML Release: Object Detection
BigML, Inc 30
Object Detection the BigML Way
• Everything is immutable, traceable, and composable


• All of your models are always available and never change


• You can always see exactly the training data that generated
the model


• Everything is available via the API and downloadable


• Image dataset creation and bounding box labeling can be
done via the API


• Models are live via API endpoint as soon as they are created


• Models can be downloaded and run locally
Democratizing Object Detection

Weitere ähnliche Inhalte

Was ist angesagt?

Azure Machine Learning
Azure Machine LearningAzure Machine Learning
Azure Machine LearningMostafa
 
Introducing DocumentDB
Introducing DocumentDB Introducing DocumentDB
Introducing DocumentDB James Serra
 
Power bi introduction
Power bi introductionPower bi introduction
Power bi introductionBishwadeb Dey
 
Business Intelligence (BI) and Data Management Basics
Business Intelligence (BI) and Data Management  Basics Business Intelligence (BI) and Data Management  Basics
Business Intelligence (BI) and Data Management Basics amorshed
 
Data Visualization Using PowerBI.pdf
Data Visualization Using PowerBI.pdfData Visualization Using PowerBI.pdf
Data Visualization Using PowerBI.pdfR20BBA040
 
A deep dive session on Tableau
A deep dive session on TableauA deep dive session on Tableau
A deep dive session on TableauVisual_BI
 
powerbi-presentation.pptx
powerbi-presentation.pptxpowerbi-presentation.pptx
powerbi-presentation.pptxAyushi716489
 
Introduction to power BI
Introduction to power BIIntroduction to power BI
Introduction to power BIRamar Bose
 
Building an Effective Data Warehouse Architecture
Building an Effective Data Warehouse ArchitectureBuilding an Effective Data Warehouse Architecture
Building an Effective Data Warehouse ArchitectureJames Serra
 
Working with Microsoft Power Business Inteligence Tools - Presented by Atidan
Working with Microsoft Power Business Inteligence Tools - Presented by AtidanWorking with Microsoft Power Business Inteligence Tools - Presented by Atidan
Working with Microsoft Power Business Inteligence Tools - Presented by AtidanDavid J Rosenthal
 
Snowflake: The Good, the Bad, and the Ugly
Snowflake: The Good, the Bad, and the UglySnowflake: The Good, the Bad, and the Ugly
Snowflake: The Good, the Bad, and the UglyTyler Wishnoff
 
Microsoft Power BI Technical Overview
Microsoft Power BI Technical OverviewMicrosoft Power BI Technical Overview
Microsoft Power BI Technical OverviewDavid J Rosenthal
 
Tag.bio: Self Service Data Mesh Platform
Tag.bio: Self Service Data Mesh PlatformTag.bio: Self Service Data Mesh Platform
Tag.bio: Self Service Data Mesh PlatformSanjay Padhi, Ph.D
 
Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...
Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...
Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...Jouko Nyholm
 

Was ist angesagt? (20)

Power BI
Power BIPower BI
Power BI
 
Azure Machine Learning
Azure Machine LearningAzure Machine Learning
Azure Machine Learning
 
Introducing DocumentDB
Introducing DocumentDB Introducing DocumentDB
Introducing DocumentDB
 
Power bi introduction
Power bi introductionPower bi introduction
Power bi introduction
 
Tableau vs PowerBI
Tableau vs PowerBITableau vs PowerBI
Tableau vs PowerBI
 
Power bi software
Power bi softwarePower bi software
Power bi software
 
Power bi overview
Power bi overview Power bi overview
Power bi overview
 
Business Intelligence (BI) and Data Management Basics
Business Intelligence (BI) and Data Management  Basics Business Intelligence (BI) and Data Management  Basics
Business Intelligence (BI) and Data Management Basics
 
RPA Uipath Presentation.pptx
RPA Uipath Presentation.pptxRPA Uipath Presentation.pptx
RPA Uipath Presentation.pptx
 
Data Visualization Using PowerBI.pdf
Data Visualization Using PowerBI.pdfData Visualization Using PowerBI.pdf
Data Visualization Using PowerBI.pdf
 
A deep dive session on Tableau
A deep dive session on TableauA deep dive session on Tableau
A deep dive session on Tableau
 
powerbi-presentation.pptx
powerbi-presentation.pptxpowerbi-presentation.pptx
powerbi-presentation.pptx
 
Introduction to power BI
Introduction to power BIIntroduction to power BI
Introduction to power BI
 
Building an Effective Data Warehouse Architecture
Building an Effective Data Warehouse ArchitectureBuilding an Effective Data Warehouse Architecture
Building an Effective Data Warehouse Architecture
 
Working with Microsoft Power Business Inteligence Tools - Presented by Atidan
Working with Microsoft Power Business Inteligence Tools - Presented by AtidanWorking with Microsoft Power Business Inteligence Tools - Presented by Atidan
Working with Microsoft Power Business Inteligence Tools - Presented by Atidan
 
Snowflake: The Good, the Bad, and the Ugly
Snowflake: The Good, the Bad, and the UglySnowflake: The Good, the Bad, and the Ugly
Snowflake: The Good, the Bad, and the Ugly
 
Power BI Overview
Power BI Overview Power BI Overview
Power BI Overview
 
Microsoft Power BI Technical Overview
Microsoft Power BI Technical OverviewMicrosoft Power BI Technical Overview
Microsoft Power BI Technical Overview
 
Tag.bio: Self Service Data Mesh Platform
Tag.bio: Self Service Data Mesh PlatformTag.bio: Self Service Data Mesh Platform
Tag.bio: Self Service Data Mesh Platform
 
Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...
Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...
Power BI Governance and Development Best Practices - Presentation at #MSBIFI ...
 

Ähnlich wie Democratizing Object Detection

AISF19 - Unleash Computer Vision at the Edge
AISF19 - Unleash Computer Vision at the EdgeAISF19 - Unleash Computer Vision at the Edge
AISF19 - Unleash Computer Vision at the EdgeBill Liu
 
Detection of a user-defined object in an image using feature extraction- Trai...
Detection of a user-defined object in an image using feature extraction- Trai...Detection of a user-defined object in an image using feature extraction- Trai...
Detection of a user-defined object in an image using feature extraction- Trai...IRJET Journal
 
BSSML17 - Feature Engineering
BSSML17 - Feature EngineeringBSSML17 - Feature Engineering
BSSML17 - Feature EngineeringBigML, Inc
 
AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...
AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...
AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...Amazon Web Services
 
MLSEV. Machine Learning: Business Perspective
MLSEV. Machine Learning: Business PerspectiveMLSEV. Machine Learning: Business Perspective
MLSEV. Machine Learning: Business PerspectiveBigML, Inc
 
DutchMLSchool. ML Business Perspective
DutchMLSchool. ML Business PerspectiveDutchMLSchool. ML Business Perspective
DutchMLSchool. ML Business PerspectiveBigML, Inc
 
BSSML17 - Deepnets
BSSML17 - DeepnetsBSSML17 - Deepnets
BSSML17 - DeepnetsBigML, Inc
 
Explainable AI in Industry (FAT* 2020 Tutorial)
Explainable AI in Industry (FAT* 2020 Tutorial)Explainable AI in Industry (FAT* 2020 Tutorial)
Explainable AI in Industry (FAT* 2020 Tutorial)Krishnaram Kenthapadi
 
VSSML18. Clustering and Latent Dirichlet Allocation
VSSML18. Clustering and Latent Dirichlet AllocationVSSML18. Clustering and Latent Dirichlet Allocation
VSSML18. Clustering and Latent Dirichlet AllocationBigML, Inc
 
Webinar: Machine Learning para Microcontroladores
Webinar: Machine Learning para MicrocontroladoresWebinar: Machine Learning para Microcontroladores
Webinar: Machine Learning para MicrocontroladoresEmbarcados
 
لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...
لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...
لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...Egyptian Engineers Association
 
VSSML18. Advanced WhizzML Workflows
VSSML18. Advanced WhizzML WorkflowsVSSML18. Advanced WhizzML Workflows
VSSML18. Advanced WhizzML WorkflowsBigML, Inc
 
Microsoft COCO: Common Objects in Context
Microsoft COCO: Common Objects in Context Microsoft COCO: Common Objects in Context
Microsoft COCO: Common Objects in Context KhalidKhan412
 
Partial Object Detection in Inclined Weather Conditions
Partial Object Detection in Inclined Weather ConditionsPartial Object Detection in Inclined Weather Conditions
Partial Object Detection in Inclined Weather ConditionsIRJET Journal
 
DutchMLSchool. Logistic Regression, Deepnets, Time Series
DutchMLSchool. Logistic Regression, Deepnets, Time SeriesDutchMLSchool. Logistic Regression, Deepnets, Time Series
DutchMLSchool. Logistic Regression, Deepnets, Time SeriesBigML, Inc
 
Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...
Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...
Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...Alok Singh
 
The Incredible Disappearing Data Scientist
The Incredible Disappearing Data ScientistThe Incredible Disappearing Data Scientist
The Incredible Disappearing Data ScientistRebecca Bilbro
 

Ähnlich wie Democratizing Object Detection (20)

AISF19 - Unleash Computer Vision at the Edge
AISF19 - Unleash Computer Vision at the EdgeAISF19 - Unleash Computer Vision at the Edge
AISF19 - Unleash Computer Vision at the Edge
 
Detection of a user-defined object in an image using feature extraction- Trai...
Detection of a user-defined object in an image using feature extraction- Trai...Detection of a user-defined object in an image using feature extraction- Trai...
Detection of a user-defined object in an image using feature extraction- Trai...
 
BSSML17 - Feature Engineering
BSSML17 - Feature EngineeringBSSML17 - Feature Engineering
BSSML17 - Feature Engineering
 
Ai use cases
Ai use casesAi use cases
Ai use cases
 
AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...
AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...
AWS re:Invent 2016: Transforming Industrial Processes with Deep Learning (MAC...
 
MLSEV. Machine Learning: Business Perspective
MLSEV. Machine Learning: Business PerspectiveMLSEV. Machine Learning: Business Perspective
MLSEV. Machine Learning: Business Perspective
 
DutchMLSchool. ML Business Perspective
DutchMLSchool. ML Business PerspectiveDutchMLSchool. ML Business Perspective
DutchMLSchool. ML Business Perspective
 
Debugging AI
Debugging AIDebugging AI
Debugging AI
 
BSSML17 - Deepnets
BSSML17 - DeepnetsBSSML17 - Deepnets
BSSML17 - Deepnets
 
Explainable AI in Industry (FAT* 2020 Tutorial)
Explainable AI in Industry (FAT* 2020 Tutorial)Explainable AI in Industry (FAT* 2020 Tutorial)
Explainable AI in Industry (FAT* 2020 Tutorial)
 
VSSML18. Clustering and Latent Dirichlet Allocation
VSSML18. Clustering and Latent Dirichlet AllocationVSSML18. Clustering and Latent Dirichlet Allocation
VSSML18. Clustering and Latent Dirichlet Allocation
 
Webinar: Machine Learning para Microcontroladores
Webinar: Machine Learning para MicrocontroladoresWebinar: Machine Learning para Microcontroladores
Webinar: Machine Learning para Microcontroladores
 
لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...
لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...
لموعد الإثنين 03 يناير 2022 143 مبادرة #تواصل_تطوير المحاضرة ال 143 من المباد...
 
VSSML18. Advanced WhizzML Workflows
VSSML18. Advanced WhizzML WorkflowsVSSML18. Advanced WhizzML Workflows
VSSML18. Advanced WhizzML Workflows
 
Microsoft COCO: Common Objects in Context
Microsoft COCO: Common Objects in Context Microsoft COCO: Common Objects in Context
Microsoft COCO: Common Objects in Context
 
Partial Object Detection in Inclined Weather Conditions
Partial Object Detection in Inclined Weather ConditionsPartial Object Detection in Inclined Weather Conditions
Partial Object Detection in Inclined Weather Conditions
 
DutchMLSchool. Logistic Regression, Deepnets, Time Series
DutchMLSchool. Logistic Regression, Deepnets, Time SeriesDutchMLSchool. Logistic Regression, Deepnets, Time Series
DutchMLSchool. Logistic Regression, Deepnets, Time Series
 
Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...
Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...
Big Data Spain 2018: How to build Weighted XGBoost ML model for Imbalance dat...
 
The Incredible Disappearing Data Scientist
The Incredible Disappearing Data ScientistThe Incredible Disappearing Data Scientist
The Incredible Disappearing Data Scientist
 
OpenCV+Android.pptx
OpenCV+Android.pptxOpenCV+Android.pptx
OpenCV+Android.pptx
 

Mehr von BigML, Inc

Digital Transformation and Process Optimization in Manufacturing
Digital Transformation and Process Optimization in ManufacturingDigital Transformation and Process Optimization in Manufacturing
Digital Transformation and Process Optimization in ManufacturingBigML, Inc
 
DutchMLSchool 2022 - Automation
DutchMLSchool 2022 - AutomationDutchMLSchool 2022 - Automation
DutchMLSchool 2022 - AutomationBigML, Inc
 
DutchMLSchool 2022 - ML for AML Compliance
DutchMLSchool 2022 - ML for AML ComplianceDutchMLSchool 2022 - ML for AML Compliance
DutchMLSchool 2022 - ML for AML ComplianceBigML, Inc
 
DutchMLSchool 2022 - Multi Perspective Anomalies
DutchMLSchool 2022 - Multi Perspective AnomaliesDutchMLSchool 2022 - Multi Perspective Anomalies
DutchMLSchool 2022 - Multi Perspective AnomaliesBigML, Inc
 
DutchMLSchool 2022 - My First Anomaly Detector
DutchMLSchool 2022 - My First Anomaly Detector DutchMLSchool 2022 - My First Anomaly Detector
DutchMLSchool 2022 - My First Anomaly Detector BigML, Inc
 
DutchMLSchool 2022 - Anomaly Detection
DutchMLSchool 2022 - Anomaly DetectionDutchMLSchool 2022 - Anomaly Detection
DutchMLSchool 2022 - Anomaly DetectionBigML, Inc
 
DutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in MLDutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in MLBigML, Inc
 
DutchMLSchool 2022 - End-to-End ML
DutchMLSchool 2022 - End-to-End MLDutchMLSchool 2022 - End-to-End ML
DutchMLSchool 2022 - End-to-End MLBigML, Inc
 
DutchMLSchool 2022 - A Data-Driven Company
DutchMLSchool 2022 - A Data-Driven CompanyDutchMLSchool 2022 - A Data-Driven Company
DutchMLSchool 2022 - A Data-Driven CompanyBigML, Inc
 
DutchMLSchool 2022 - ML in the Legal Sector
DutchMLSchool 2022 - ML in the Legal SectorDutchMLSchool 2022 - ML in the Legal Sector
DutchMLSchool 2022 - ML in the Legal SectorBigML, Inc
 
DutchMLSchool 2022 - Smart Safe Stadiums
DutchMLSchool 2022 - Smart Safe StadiumsDutchMLSchool 2022 - Smart Safe Stadiums
DutchMLSchool 2022 - Smart Safe StadiumsBigML, Inc
 
DutchMLSchool 2022 - Process Optimization in Manufacturing Plants
DutchMLSchool 2022 - Process Optimization in Manufacturing PlantsDutchMLSchool 2022 - Process Optimization in Manufacturing Plants
DutchMLSchool 2022 - Process Optimization in Manufacturing PlantsBigML, Inc
 
DutchMLSchool 2022 - Anomaly Detection at Scale
DutchMLSchool 2022 - Anomaly Detection at ScaleDutchMLSchool 2022 - Anomaly Detection at Scale
DutchMLSchool 2022 - Anomaly Detection at ScaleBigML, Inc
 
DutchMLSchool 2022 - Citizen Development in AI
DutchMLSchool 2022 - Citizen Development in AIDutchMLSchool 2022 - Citizen Development in AI
DutchMLSchool 2022 - Citizen Development in AIBigML, Inc
 
Machine Learning in Retail: Know Your Customers' Customer. See Your Future
Machine Learning in Retail: Know Your Customers' Customer. See Your FutureMachine Learning in Retail: Know Your Customers' Customer. See Your Future
Machine Learning in Retail: Know Your Customers' Customer. See Your FutureBigML, Inc
 
Machine Learning in Retail: ML in the Retail Sector
Machine Learning in Retail: ML in the Retail SectorMachine Learning in Retail: ML in the Retail Sector
Machine Learning in Retail: ML in the Retail SectorBigML, Inc
 
ML in GRC: Machine Learning in Legal Automation, How to Trust a Lawyerbot
ML in GRC: Machine Learning in Legal Automation, How to Trust a LawyerbotML in GRC: Machine Learning in Legal Automation, How to Trust a Lawyerbot
ML in GRC: Machine Learning in Legal Automation, How to Trust a LawyerbotBigML, Inc
 
ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...
ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...
ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...BigML, Inc
 
ML in GRC: Cybersecurity versus Governance, Risk Management, and Compliance
ML in GRC: Cybersecurity versus Governance, Risk Management, and ComplianceML in GRC: Cybersecurity versus Governance, Risk Management, and Compliance
ML in GRC: Cybersecurity versus Governance, Risk Management, and ComplianceBigML, Inc
 
Intelligent Mobility: Machine Learning in the Mobility Industry
Intelligent Mobility: Machine Learning in the Mobility IndustryIntelligent Mobility: Machine Learning in the Mobility Industry
Intelligent Mobility: Machine Learning in the Mobility IndustryBigML, Inc
 

Mehr von BigML, Inc (20)

Digital Transformation and Process Optimization in Manufacturing
Digital Transformation and Process Optimization in ManufacturingDigital Transformation and Process Optimization in Manufacturing
Digital Transformation and Process Optimization in Manufacturing
 
DutchMLSchool 2022 - Automation
DutchMLSchool 2022 - AutomationDutchMLSchool 2022 - Automation
DutchMLSchool 2022 - Automation
 
DutchMLSchool 2022 - ML for AML Compliance
DutchMLSchool 2022 - ML for AML ComplianceDutchMLSchool 2022 - ML for AML Compliance
DutchMLSchool 2022 - ML for AML Compliance
 
DutchMLSchool 2022 - Multi Perspective Anomalies
DutchMLSchool 2022 - Multi Perspective AnomaliesDutchMLSchool 2022 - Multi Perspective Anomalies
DutchMLSchool 2022 - Multi Perspective Anomalies
 
DutchMLSchool 2022 - My First Anomaly Detector
DutchMLSchool 2022 - My First Anomaly Detector DutchMLSchool 2022 - My First Anomaly Detector
DutchMLSchool 2022 - My First Anomaly Detector
 
DutchMLSchool 2022 - Anomaly Detection
DutchMLSchool 2022 - Anomaly DetectionDutchMLSchool 2022 - Anomaly Detection
DutchMLSchool 2022 - Anomaly Detection
 
DutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in MLDutchMLSchool 2022 - History and Developments in ML
DutchMLSchool 2022 - History and Developments in ML
 
DutchMLSchool 2022 - End-to-End ML
DutchMLSchool 2022 - End-to-End MLDutchMLSchool 2022 - End-to-End ML
DutchMLSchool 2022 - End-to-End ML
 
DutchMLSchool 2022 - A Data-Driven Company
DutchMLSchool 2022 - A Data-Driven CompanyDutchMLSchool 2022 - A Data-Driven Company
DutchMLSchool 2022 - A Data-Driven Company
 
DutchMLSchool 2022 - ML in the Legal Sector
DutchMLSchool 2022 - ML in the Legal SectorDutchMLSchool 2022 - ML in the Legal Sector
DutchMLSchool 2022 - ML in the Legal Sector
 
DutchMLSchool 2022 - Smart Safe Stadiums
DutchMLSchool 2022 - Smart Safe StadiumsDutchMLSchool 2022 - Smart Safe Stadiums
DutchMLSchool 2022 - Smart Safe Stadiums
 
DutchMLSchool 2022 - Process Optimization in Manufacturing Plants
DutchMLSchool 2022 - Process Optimization in Manufacturing PlantsDutchMLSchool 2022 - Process Optimization in Manufacturing Plants
DutchMLSchool 2022 - Process Optimization in Manufacturing Plants
 
DutchMLSchool 2022 - Anomaly Detection at Scale
DutchMLSchool 2022 - Anomaly Detection at ScaleDutchMLSchool 2022 - Anomaly Detection at Scale
DutchMLSchool 2022 - Anomaly Detection at Scale
 
DutchMLSchool 2022 - Citizen Development in AI
DutchMLSchool 2022 - Citizen Development in AIDutchMLSchool 2022 - Citizen Development in AI
DutchMLSchool 2022 - Citizen Development in AI
 
Machine Learning in Retail: Know Your Customers' Customer. See Your Future
Machine Learning in Retail: Know Your Customers' Customer. See Your FutureMachine Learning in Retail: Know Your Customers' Customer. See Your Future
Machine Learning in Retail: Know Your Customers' Customer. See Your Future
 
Machine Learning in Retail: ML in the Retail Sector
Machine Learning in Retail: ML in the Retail SectorMachine Learning in Retail: ML in the Retail Sector
Machine Learning in Retail: ML in the Retail Sector
 
ML in GRC: Machine Learning in Legal Automation, How to Trust a Lawyerbot
ML in GRC: Machine Learning in Legal Automation, How to Trust a LawyerbotML in GRC: Machine Learning in Legal Automation, How to Trust a Lawyerbot
ML in GRC: Machine Learning in Legal Automation, How to Trust a Lawyerbot
 
ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...
ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...
ML in GRC: Supporting Human Decision Making for Regulatory Adherence with Mac...
 
ML in GRC: Cybersecurity versus Governance, Risk Management, and Compliance
ML in GRC: Cybersecurity versus Governance, Risk Management, and ComplianceML in GRC: Cybersecurity versus Governance, Risk Management, and Compliance
ML in GRC: Cybersecurity versus Governance, Risk Management, and Compliance
 
Intelligent Mobility: Machine Learning in the Mobility Industry
Intelligent Mobility: Machine Learning in the Mobility IndustryIntelligent Mobility: Machine Learning in the Mobility Industry
Intelligent Mobility: Machine Learning in the Mobility Industry
 

Kürzlich hochgeladen

Call me @ 9892124323 Cheap Rate Call Girls in Vashi with Real Photo 100% Secure
Call me @ 9892124323  Cheap Rate Call Girls in Vashi with Real Photo 100% SecureCall me @ 9892124323  Cheap Rate Call Girls in Vashi with Real Photo 100% Secure
Call me @ 9892124323 Cheap Rate Call Girls in Vashi with Real Photo 100% SecurePooja Nehwal
 
Al Barsha Escorts $#$ O565212860 $#$ Escort Service In Al Barsha
Al Barsha Escorts $#$ O565212860 $#$ Escort Service In Al BarshaAl Barsha Escorts $#$ O565212860 $#$ Escort Service In Al Barsha
Al Barsha Escorts $#$ O565212860 $#$ Escort Service In Al BarshaAroojKhan71
 
Smarteg dropshipping via API with DroFx.pptx
Smarteg dropshipping via API with DroFx.pptxSmarteg dropshipping via API with DroFx.pptx
Smarteg dropshipping via API with DroFx.pptxolyaivanovalion
 
April 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's AnalysisApril 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's Analysismanisha194592
 
Ravak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptxRavak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptxolyaivanovalion
 
FESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdfFESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdfMarinCaroMartnezBerg
 
Edukaciniai dropshipping via API with DroFx
Edukaciniai dropshipping via API with DroFxEdukaciniai dropshipping via API with DroFx
Edukaciniai dropshipping via API with DroFxolyaivanovalion
 
BDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort Service
BDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort ServiceBDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort Service
BDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort ServiceDelhi Call girls
 
Delhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip CallDelhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Callshivangimorya083
 
Introduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptxIntroduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptxfirstjob4
 
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...Delhi Call girls
 
Best VIP Call Girls Noida Sector 22 Call Me: 8448380779
Best VIP Call Girls Noida Sector 22 Call Me: 8448380779Best VIP Call Girls Noida Sector 22 Call Me: 8448380779
Best VIP Call Girls Noida Sector 22 Call Me: 8448380779Delhi Call girls
 
Best VIP Call Girls Noida Sector 39 Call Me: 8448380779
Best VIP Call Girls Noida Sector 39 Call Me: 8448380779Best VIP Call Girls Noida Sector 39 Call Me: 8448380779
Best VIP Call Girls Noida Sector 39 Call Me: 8448380779Delhi Call girls
 
CebaBaby dropshipping via API with DroFX.pptx
CebaBaby dropshipping via API with DroFX.pptxCebaBaby dropshipping via API with DroFX.pptx
CebaBaby dropshipping via API with DroFX.pptxolyaivanovalion
 
Halmar dropshipping via API with DroFx
Halmar  dropshipping  via API with DroFxHalmar  dropshipping  via API with DroFx
Halmar dropshipping via API with DroFxolyaivanovalion
 
Mature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptxMature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptxolyaivanovalion
 
100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptx100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptxAnupama Kate
 
Zuja dropshipping via API with DroFx.pptx
Zuja dropshipping via API with DroFx.pptxZuja dropshipping via API with DroFx.pptx
Zuja dropshipping via API with DroFx.pptxolyaivanovalion
 

Kürzlich hochgeladen (20)

Delhi 99530 vip 56974 Genuine Escort Service Call Girls in Kishangarh
Delhi 99530 vip 56974 Genuine Escort Service Call Girls in  KishangarhDelhi 99530 vip 56974 Genuine Escort Service Call Girls in  Kishangarh
Delhi 99530 vip 56974 Genuine Escort Service Call Girls in Kishangarh
 
Call me @ 9892124323 Cheap Rate Call Girls in Vashi with Real Photo 100% Secure
Call me @ 9892124323  Cheap Rate Call Girls in Vashi with Real Photo 100% SecureCall me @ 9892124323  Cheap Rate Call Girls in Vashi with Real Photo 100% Secure
Call me @ 9892124323 Cheap Rate Call Girls in Vashi with Real Photo 100% Secure
 
Al Barsha Escorts $#$ O565212860 $#$ Escort Service In Al Barsha
Al Barsha Escorts $#$ O565212860 $#$ Escort Service In Al BarshaAl Barsha Escorts $#$ O565212860 $#$ Escort Service In Al Barsha
Al Barsha Escorts $#$ O565212860 $#$ Escort Service In Al Barsha
 
Smarteg dropshipping via API with DroFx.pptx
Smarteg dropshipping via API with DroFx.pptxSmarteg dropshipping via API with DroFx.pptx
Smarteg dropshipping via API with DroFx.pptx
 
April 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's AnalysisApril 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's Analysis
 
Ravak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptxRavak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptx
 
FESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdfFESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdf
 
Edukaciniai dropshipping via API with DroFx
Edukaciniai dropshipping via API with DroFxEdukaciniai dropshipping via API with DroFx
Edukaciniai dropshipping via API with DroFx
 
Sampling (random) method and Non random.ppt
Sampling (random) method and Non random.pptSampling (random) method and Non random.ppt
Sampling (random) method and Non random.ppt
 
BDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort Service
BDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort ServiceBDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort Service
BDSM⚡Call Girls in Mandawali Delhi >༒8448380779 Escort Service
 
Delhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip CallDelhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls CP 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
 
Introduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptxIntroduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptx
 
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
 
Best VIP Call Girls Noida Sector 22 Call Me: 8448380779
Best VIP Call Girls Noida Sector 22 Call Me: 8448380779Best VIP Call Girls Noida Sector 22 Call Me: 8448380779
Best VIP Call Girls Noida Sector 22 Call Me: 8448380779
 
Best VIP Call Girls Noida Sector 39 Call Me: 8448380779
Best VIP Call Girls Noida Sector 39 Call Me: 8448380779Best VIP Call Girls Noida Sector 39 Call Me: 8448380779
Best VIP Call Girls Noida Sector 39 Call Me: 8448380779
 
CebaBaby dropshipping via API with DroFX.pptx
CebaBaby dropshipping via API with DroFX.pptxCebaBaby dropshipping via API with DroFX.pptx
CebaBaby dropshipping via API with DroFX.pptx
 
Halmar dropshipping via API with DroFx
Halmar  dropshipping  via API with DroFxHalmar  dropshipping  via API with DroFx
Halmar dropshipping via API with DroFx
 
Mature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptxMature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptx
 
100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptx100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptx
 
Zuja dropshipping via API with DroFx.pptx
Zuja dropshipping via API with DroFx.pptxZuja dropshipping via API with DroFx.pptx
Zuja dropshipping via API with DroFx.pptx
 

Democratizing Object Detection

  • 2. BigML Release: Object Detection BigML, Inc 2 Object Detection Enter questions into chat box. We will answer some via chat and others at the end of the session. QUESTIONS info@bigml.com @bigmlcom Atakan Cetinsoy VP of Predictive Applications. MODERATOR Charles Parker Ph.D. VP of Machine Learning Algorithms. SPEAKER https://bigml.com/releases/object-detection RESOURCES
  • 3. BigML Release: Object Detection BigML, Inc 3 Object Detection is a fundamental Computer Vision task that involves localizing and classifying objects-of-interest in an image. Introducing Object Detection
  • 4. BigML Release: Object Detection BigML, Inc 4 Object Detection truck car Built on the foundation of BigML Image Processing, Object Detection allows you to localize and classify objects in a fast and scalable manner. Image Classi fi cation vs. Object Detection Image Classi fi cation
  • 5. BigML Release: Object Detection BigML, Inc 5 Maturity of DL techniques Decreasing compute costs Abundance of image data High-speed processing (frames/sec.) Next-Gen Tools Object Detection - Why now?
  • 6. BigML Release: Object Detection BigML, Inc 6 Object Detection is Everywhere! Healthcare Agriculture Manufacturing Sports Transportation Retail
  • 7. BigML Release: Object Detection BigML, Inc 7 Get plate numbers and states from distinct highway toll car images Toll Video Cameras 1 Individual Car Images 2 Plate Detection 3 BigML Object Detection Plate Extraction 4 5 Characters Recognition 6 State Classi fi er BigML Object Detection Character Images Classi fi cation Image Classi fi cation 7 Automatic Payment After detecting plate number and state, automatic payment is possible Object Detection in Transportation
  • 8. BigML Release: Object Detection BigML, Inc 8 Smart Video Review Framework: Find, Group, Label, and Review Find interesting events Flags Chants Group events Experts review clusters and assign their own labels 1 2 Find anomalous events Cluster similar events Smoke bombs / Flares Label event groups 3 4 Assign labels to known events Models learn new labels from expert’s feedback Object Detection for Public Safety
  • 9. BigML Release: Object Detection BigML, Inc 9 Object Detection without BigML Image Data Labeled Image Data Labeling Tool Modeling Tool Predictions Con fi gured Machine w/ GPU Modeling Hardware or Service Trained Model LABELERS ML ENGINEER IT DEPARTMENT
  • 10. BigML Release: Object Detection BigML, Inc 10 Scaling Object Detection without BigML Image Data Labeled Image Data Trained Model Con fi gure d Machine w/ GPU Labeling Tool Modeling Tool Modeling Hardware or Predictions Con fi gure d Machine w/ GPU Con fi gure d Machine w/ GPU Con fi gure d Machine w/ GPU Con fi gure d Machine w/ GPU Con fi gured Machine w/ GPU Modeling Hardware or Modeling Hardware or Modeling Hardware or Modeling Hardware or Service MLOps / Reporting Trained Model Trained Model LABELERS ML ENGINEER IT DEPARTMENT
  • 11. BigML Release: Object Detection BigML, Inc 11 Scalable Object Detection with BigML LABELERS ML ENGINEER Image Data Predictions
  • 12. BigML Release: Object Detection BigML, Inc 12 BigML Composability ML Work fl ow Object Detection BigML Resources High-Level Algorithm 1-Click Execution
  • 13. BigML Release: Object Detection BigML, Inc 13 BigML Resource Synergies
  • 14. BigML Release: Object Detection BigML, Inc 14 Handling Images the BigML Way • ALL the tasks for generating insights from image data ON A SINGLE PLATFORM: From labeling to inference, evaluation, and predictions.  • Streamlined image dataset management with composite sources. • Automatic handling of infrastructure concerns. • Your choice: Code and no code. Solving image data-driven business problems with remarkable ease of use
  • 15. BigML Release: Object Detection BigML, Inc 15 What Object Detection Is • Simultaneously locating and classifying specific sub-areas (returned as xmin, ymin, xmax, ymax) in a given image. • Many uses, but especially in: • Manufacturing defect detection • Medical image analysis • Autonomous navigation • Security (monitoring) applications
  • 16. BigML Release: Object Detection BigML, Inc 16 What Object Detection Is • Identification of a specific area in the image that contains an object and identification of the class of the object. • The “score” given to the box is a product of the probability of both of these factors. Note that even if the model thinks there’s an object in a particular region, it can still fail if the model has trouble deducing the class of the object. • While we typically visualize the prediction as a box, this is just for convenience. • The parameters xmin, ymin, xmax, ymax could also specify an ellipse • The boundaries of the object are unlikely to be pixel-accurate, even for objects that are axis-aligned rectangles • Better to think of the returned values is “approximate location” and “approximate size”
  • 17. BigML Release: Object Detection BigML, Inc 17 What Object Detection Isn’t • Object detection tells you the class for a specific window of the image. If you want to classify the whole image, you want “image classification” (available on BigML). • Object detection does not give an idea of exactly which pixels in the image are in the detected object. This is done using one of the various flavors of “image segmentation” algorithms. • Object detection on its own cannot find specific parts of the object being detected. This is “keypoint detection”, or for articulated objects (like people), “pose estimation”. Though it is possible to label and detect parts separately with object detection.
  • 18. BigML Release: Object Detection BigML, Inc 18 New Optype: Regions Type Description Example Numeric A continuous numeric value 0,75 Categorical One of a small set of possible values Cat Datetime A String specifying a date, time, or both 4-22-2022 11:24am Text An unstructured string of free text “Hello, cats!” Image A fi le containing an image Path The path of the image fi le in the zip fi le containing it cat/image1.jpg Regions A list of fi ve-tuples, [label, xmin, ymin, xmax, ymax] [[“cat”, 0.5, 0.1, 0.6, 0.9], [“dog”, 0.2, 0.2, 0.8, 0.8]]
  • 19. BigML, Inc BigML Release: Object Detection Regions Labeling 19
  • 20. BigML Release: Object Detection BigML, Inc 20 The More Things Change . . . Training Dataset Test Dataset Zip fi le with Images Train Evaluate
  • 21. BigML, Inc BigML Release: Object Detection Training An Object Detector 21
  • 22. BigML Release: Object Detection BigML, Inc 22 Under The Hood • Break the image into a rough grid of cells. • Centered at each cell, consider a number of “bounding box proposals”. • For each box in each cell, try to predict the following things simultaneously: • Whether or not there’s a object inside. • The class of the object. • Adjustments needed for the height/width/x-center/y- center to make the box fit properly around the object. • The shapes of the proposals are learned directly from the data.
  • 23. BigML Release: Object Detection BigML, Inc 23 Some Caveats • The usual caveats about feeding images to machine learning models still apply. • Example: Blurring. • Example: Negative images. • The bounding box proposals are learned from the sizes given in your data. • The boxes consider the entire image when making predictions, not just the content of the box proposal itself.
  • 24. BigML Release: Object Detection BigML, Inc 24 What if we Need More Power? • For the public deployment, we use a small object detection architecture which performs reasonably well on most problems. • There are, however, other architectures that can give a significant bump in accuracy. • They need significantly more data to train properly. • They are slower for both training and testing. • We don’t make these available on our public deployment, but contact us about these options in private deployments.
  • 25. BigML, Inc BigML Release: Object Detection Making Predictions with Your Detector 25
  • 26. BigML Release: Object Detection BigML, Inc 26 Evaluations - A Little Different • In classification problems, we can construct a “confusion matrix” with true positives, true negatives, false positives, and false negatives. • For object detection, we don’t typically count “true negatives” as we don’t give the model credit for predicting “nothing” for the vast majority of locations, which would dominate metrics that consider them. • So we must rely on metrics that don’t involve true negatives, which are based on precision, recall, and their combinations (e.g., average precision; the precision averaged over several levels of recall). • We can still change the score threshold for the model, and see how that impacts the count of true positives, false positives, and false negatives.
  • 27. BigML Release: Object Detection BigML, Inc 27 What is a “Correct” Prediction? • Clearly, if the boxes align exactly, the prediction is correct, but they usually don’t. • For each possible match, we compute a quantity called “Intersection over Union” or IOU. • If the IOU is greater than some threshold, we consider the prediction correct. • We do this for several thresholds in the evaluation.
  • 28. BigML, Inc BigML Release: Object Detection Object Detection Evaluation 28
  • 29. BigML Release: Object Detection BigML, Inc Object Detection Model Classi fi er 29 A Simple Work fl ow Anomaly Detector Object 1 Object 2 Object Classes Anomaly Scores • What if you have a basic list of things you would like to detect, but you know it’s not exhaustive? • This might cause low scores for rare / never seen objects. • Solution: Train an object detector to detect a generic “thing”. • Crop out the detected objects, then passed the crops to both an anomaly detector and a classifier. • If it’s a known object, you’ll get a high- confidence class. If it’s a strange object, you will get a high anomaly score.
  • 30. BigML Release: Object Detection BigML, Inc 30 Object Detection the BigML Way • Everything is immutable, traceable, and composable • All of your models are always available and never change • You can always see exactly the training data that generated the model • Everything is available via the API and downloadable • Image dataset creation and bounding box labeling can be done via the API • Models are live via API endpoint as soon as they are created • Models can be downloaded and run locally