SlideShare une entreprise Scribd logo

Face Recognition: From Scratch To Hatch

•Télécharger en tant que PPTX, PDF•
•
E
Eduard Tyantov

face detection & recognition: metric learning (triplet, center, angular softmax losses), TensoRT, tips & tricks.

Internet
1  sur  100
Face Recognition: From Scratch To Hatch Tyantov Eduard, Mail.ru Group
Face Recognition in Cloud@Mail.ru Users upload photos to Cloud Backend identifies persons on photos, tags and show clusters
??? You Your Ex-Girlfriend Social networks
Convolutional neural networks, briefly
edges object parts (combination of edges) object models
Face Detection
Face detection
Auxiliary task: facial landmarks – Face alignment: rotation – Goal: make it easier for Face Recognition Rotate
Train Datasets Wider – 32k images – 494k faces Celeba – 200k images, 10k persons – Landmarks, 40 binary attributes
Test Dataset: FDDB Face Detection Data Set and Benchmark – 2845 images – 5171 faces
Old school: Viola-Jones Haar Feature-based Cascade Classifiers Haar-like features eyes darker nose lighter Examples
Viola-Jones algorithm: training for each patch AdaBoost ensemble features 160k valuable 6k weighted sum Face or NotDataset
Viola-Jones algorithm: inference for each patch Stages Face Yes Yes Stage 1 Stage 2 Stage N Optimization – Features are grouped into stages – If a patch fails any stage => discard
Viola-Jones results OpenCV implementation – Fast: ~100ms on CPU FDDB results 0.45 – Not accurate
1. Pre-trained network: extracting features New school: Region-based Convolutional Networks Faster RCNN, algorithm 1 CNN 2 RPN 4 Classifier Face ? 3 Roi-pooling Feature Maps 3 2. Region proposal network3. RoI-pooling: extract corresponding tensor4. Classifier: classes and the bounding box
Comparison: Viola-Jones vs R-FCN Results – 92% accuracy (R-FCN) FDDB results Viola-Jones (opencv) 0.45 HOG (dlib) 0.7 R-FCN 0.92 – 40ms on GPU (slow)
Face detection: how fast We need faster solution at the same accuracy! Target: < 10ms
Alternative: MTCNN 1 Different scales 2 Proposal CNN 3 Refine CNN 4 Output CNN Cascade of 3 CNN 1. Resize to different scales2. Proposal -> candidates + b-boxes 3. Refine -> calibration4. Output -> b-boxes + landmarks
Comparison: MTCNN vs R-FCN MTCNN + Faster + Landmarks - Less accurate - No batch processing Model GPU Inference FDDB Precision (100 errors) R-FCN 40 ms 92% MTCNN 17 ms 90%
TensorRT
What is TensorRT NVIDIA TensorRT is a high-performance deep learning inference optimizer Features – Improves performance for complex networks – FP16 & INT8 support – Effective at small batch-sizes
TensorRT: layer optimizations 2. Horizontal fusion 3. Concat elision 1. Vertical layer fusion
TensorRT: downsides 1. Caffe + TensorFlow supported 2. Fixed input/batch size 3. Basic layers support
Batch processing Problem Image size is fixed, but MTCNN works at different scales Solution Pyramid on a single image
Batch processing Results – Single run – Enables batch processing Model Inference ms MTCNN (Caffe, python) 17 MTCNN (Caffe, C++) 12.7 + batch 10.7
TensorRT: layers Problem No PReLU layer => default pre-trained model can’t be used Retrained with ReLU from scratch Model GPU Inference ms FDDB Precision (100 errors) MTCNN, batch 10.7 90% +Tensor RT 8.8 91.2% -20%
Face detection: inference Target: < 10 ms Result: 8.8 ms Ingredients 1. MTCNN 2. Batch processing 3. TensorRT
Face Recognition
Face recognition task – Goal – to compare faces Latent SpaceCNN Embedding close distant Unseen – How? To learn metric – To enable Zero-shot learning
Training set: MSCeleb – Top 100k celebrities – 10 Million images, 100 per person – Noisy: constructed by leveraging public search engines
Small test dataset: LFW Labeled Faces in the Wild Home – 13k images from the web – 1680 persons have >= 2 photos
Large test dataset: Megaface – Identification under up to 1 million “distractors” – 530 people to find
Megaface leaderboard ~80%
Metric Learning
Classification CNN Embedding Classify – Train CNN to predict classes – Pray for a good latent space close distant
Softmax – Learned features only separable but not discriminative – The resulting features are not sufficiently effective close
We need metric learning – Tightness of the cluster – Discriminative features
Triplet loss Features – Identity -> single point – Enforces a margin between persons Anchor Positive Negativepositive + α < negative minimize maximize
Choosing triplets Crucial problem How to choose triplets ? Useful triplets = hardest errors Pick all positive Too easy Hard enough Solution Hard-mining within a large mini-batch (>1000)
Choosing triplets: trap
Choosing triplets: trap Anchor Positive positive ~ negative minimize maximize Negative
Choosing triplets: trap Instead
Choosing triplets: trap Selecting hardest negative may lead to the collapse early in training
Choosing triplets: semi-hard Pick all positive Too easy Semi-hard Too hard positive < negative < positive + α
Triplet loss: summary Overview – Requires large batches, margin tuning – Slow convergence Opensource Code – Openface (Torch) • suboptimal implementation – Facenet, not original (TensorFlow) LFW, % Megaface Openface (Torch) 92 - Our (Torch) 99.35 65 Google’s Facenet 99.63 70.5
Center loss Idea: pull the points to class centroids
Center loss: structure – Without classification loss – collapses CNN Embedding Classify Softmax Loss λ Center Loss Pull – Final loss = Softmax loss + λ Center loss
Center Loss: different lambdas λ = 10-7
Center Loss: different lambdas λ = 10-6
Center Loss: different lambdas λ = 10-5
Center loss: summary Overview – Intra-class compactness and inter-class separability – Good performance at several other tasks Opensource Code – Caffe (original, Megaface - 65%) LFW, % Megaface Triplet Loss 99.35 65 Center Loss (Torch, ours) 99.60 71.7
Tricks: augmentation Test time augmentation – Flip image Embedding Flipped Embedding Final Embedding Average – Average embeddings – Compute 2 embeddings
Tricks: alignment Rotation Kabsch algorithm - the optimal rotation matrix that minimizes the RMSD LFW, % Megaface Center Loss 99.6 71.7 Center Loss + Tricks 99.68 73
Shades on At one point we used shades augmentation How to – Construct several sunglass textures – Place them using landmarks
Tricks: adding Eye loss CNN Embedding Person Softmax Loss Center Loss – We can force CNN to learn specific discriminative features – For celebrities eye colors are available in the Internet Eye Loss Eye Color
Eye loss: summary *Adding simple features doesn’t help, i.e gender LFW, % Megaface Center Loss + Tricks 99.68 73 Center Loss + Eye 99.68 73.5
Angular Softmax ||X||= 1 On sphere Angle discriminates ||W||= 1 b=0
Enforce larger angle Angular Softmax
Angular Softmax: different «m» m=1 m=3
Angular softmax: summary Overview – As describes in the paper: doesn’t work at all – Works using sum of losses (m=1,N) over training • only on small datasets! LFW, % Megaface Center Loss 99.6 73 Center Loss + Eye 99.68 73.5 A-Softmax (Torch) 99.68 74.2 Opensource Code – Caffe (original) – Slight modification of the loss yields 74.2%
Metric learning: summary Softmax < Triplet < Center < A-Softmax A-Softmax – With bells and whistles better than center loss Center loss Overall – Rule of thumb: use Center loss – Metric learning may improve classification performance
Fighting errors
Errors after MSCeleb: children Problem Children all look alike Result Embeddings are almost single point in the space
Errors after MSCeleb: asian Problem Face Recognition’s intolerant to Asians Reason Dataset doesn’t contain enough photos of these categories
How to fix these errors ? It’s all about data, we need diverse dataset! Natural choice – avatars of social networks
A way to construct dataset Face Detection Pick largest Face Recognition+ Clustering Cleaning algorithm 1. Face detection2. Face recognition -> embeddings3. Hierarchical clustering algorithm4. Pick the largest cluster as a personIterate after each model improvement
MSCeleb dataset’s errors MSCeleb is constructed by leveraging search engines Joe Eszterhas Joe Eszterhas and Mel Gibson public confrontation leads to the error Mel Gibson =
MSCeleb dataset’s errors Female + Male
MSCeleb dataset’s errors Asia Mix
MSCeleb dataset’s errors Dataset has been shrinked from 100k to 46k celebrities Random search engine Corrected
Results on new datasets Datasets – Train: • MSCeleb (46k) • VK-train (150k) A-Softmax on dataset Megaface MSCeleb 74.2 MSCeleb cleaned 76.2 – Test • MegaVK • Sets for children and asians MegaVK 58.4 60 + VK 77.5 87.5
Ensemble Final Embedding Concat CNN-2 CNN-1 A-Softmax model Megaface MegaVK Best single 77.5 87.5 Ensemble of 2 79.6 88.6
Workaround still … Children are still challenge for the model
Workaround Algorithm 1. Construct dataset with children 2. Compute average embedding 3. Every point inside the sphere – a child 4. Tighten distance threshold there Results This allows softening the overall threshold
How to handle big dataset It seems we can add more data infinitely, but no. Problems – Memory consumption (Softmax) – Computational costs – A lot of noise in gradients
Softmax Approximation Algorithm 1. Perform K-Means clustering using current FR model Dataset K-Means Children Women Men Smaller sets
Softmax Approximation Algorithm 1. Perform K-Means clustering using current FR model CNN Embedding Predict cluster Predict person MenPerson Softmax 2. Two Softmax heads: 1. Predicts cluster label 2. Class within the true cluster Cluster Softmax Men
Softmax Approximation Pros Push Push Harder negative 1. Prevents fusing of the clusters 2. Does hard-negative mining 3. Clusters can be specified • Children • Asian
Fighting errors on production
Errors: blur Problem • Detector yields blurry photos • Recognition forms «blurry clusters» Solution Laplacian – 2nd order derivative of the image
Laplacian in action Low variance High variance
Errors: body parts Detection mistakes form clusters
Errors: diagrams & mushrooms
Fixing trash clusters There is similarity between “no faces”! CNN Embedding Embedding No «firing» features
Fixing trash clusters «Trash» has small norm Faces Trash Softmax loss Motivation Softmax encourage big embedding’s norms Results – ROC AUC 97% – Better then Laplacian for blurry
Spectacular results
Fun: new governors Recently appointed governors are almost twins, but FR distinguishes them Dmitriy Gleb
Over years Face recognition algorithm captures similarity across years Although we didn’t focus on the problem
Over years
Summary 1. Use TensorRT to speed up inference 2. Metric learning: use Center loss by default 3. Clean your data thoroughly 4. Understanding CNN helps to fight errors
Thanks! Questions?
Auxiliary
Best avatar Problem How to pick an avatar for a person ? Solution Train model to predict awesomeness of photo
Predicting awesomeness: how to approach Social networks – not only photos, but likes too
Predicting awesomeness: dataset Awesomeness (A) = likes/audience A=18% A=27% A=75%
Results – Mean Aveage Precision @5: 25% – Data and metric are noisy => human evaluation Predicting awesomeness: summary High score Low score
Predicting awesomeness: incorporating into FR One more branch in Face Recognition CNN Small overhead awesomeness embedding face
Appendix
Histogram loss Idea – Compute similarities positive and negative pairs – Maximize a probability that a randomly sampled positive pair has smaller similarity than negative one Loss = the integral of the product between the negative distribution and the cumulative density function for the positive distribution (shown with a dashed line)
Histogram loss Results – Got nice distributions – Doesn’t improve the triplet results – LFW on 97.7%

Recommandé

PR 127: FaceNet
PR 127: FaceNetPR 127: FaceNet
PR 127: FaceNet
Taeoh Kim
 
Image classification using CNN
Image classification using CNNImage classification using CNN
Image classification using CNN
Noura Hussein
 
Deep Learning for Computer Vision: Data Augmentation (UPC 2016)
Deep Learning for Computer Vision: Data Augmentation (UPC 2016)Deep Learning for Computer Vision: Data Augmentation (UPC 2016)
Deep Learning for Computer Vision: Data Augmentation (UPC 2016)
Universitat Politècnica de Catalunya
 
210110 deformable detr
210110 deformable detr210110 deformable detr
210110 deformable detr
taeseon ryu
 
AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)
AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)
AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)
Fellowship at Vodafone FutureLab
 
Image Recognition Expert System based on deep learning
Image Recognition Expert System based on deep learningImage Recognition Expert System based on deep learning
Image Recognition Expert System based on deep learning
PRATHAMESH REGE
 
ViT (Vision Transformer) Review [CDM]
ViT (Vision Transformer) Review [CDM]ViT (Vision Transformer) Review [CDM]
ViT (Vision Transformer) Review [CDM]
Dongmin Choi
 
Object Detection using Deep Neural Networks
Object Detection using Deep Neural NetworksObject Detection using Deep Neural Networks
Object Detection using Deep Neural Networks
Usman Qayyum
 

Recommandé

PR 127: FaceNet
PR 127: FaceNetPR 127: FaceNet
PR 127: FaceNet
Taeoh Kim
 
Image classification using CNN
Image classification using CNNImage classification using CNN
Image classification using CNN
Noura Hussein
 
Deep Learning for Computer Vision: Data Augmentation (UPC 2016)
Deep Learning for Computer Vision: Data Augmentation (UPC 2016)Deep Learning for Computer Vision: Data Augmentation (UPC 2016)
Deep Learning for Computer Vision: Data Augmentation (UPC 2016)
Universitat Politècnica de Catalunya
 
210110 deformable detr
210110 deformable detr210110 deformable detr
210110 deformable detr
taeseon ryu
 
AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)
AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)
AlexNet(ImageNet Classification with Deep Convolutional Neural Networks)
Fellowship at Vodafone FutureLab
 
Image Recognition Expert System based on deep learning
Image Recognition Expert System based on deep learningImage Recognition Expert System based on deep learning
Image Recognition Expert System based on deep learning
PRATHAMESH REGE
 
ViT (Vision Transformer) Review [CDM]
ViT (Vision Transformer) Review [CDM]ViT (Vision Transformer) Review [CDM]
ViT (Vision Transformer) Review [CDM]
Dongmin Choi
 
Object Detection using Deep Neural Networks
Object Detection using Deep Neural NetworksObject Detection using Deep Neural Networks
Object Detection using Deep Neural Networks
Usman Qayyum
 
A brief introduction to recent segmentation methods
A brief introduction to recent segmentation methodsA brief introduction to recent segmentation methods
A brief introduction to recent segmentation methods
Shunta Saito
 
Breast Cancer Detection with Convolutional Neural Networks (CNN)
Breast Cancer Detection with Convolutional Neural Networks (CNN)Breast Cancer Detection with Convolutional Neural Networks (CNN)
Breast Cancer Detection with Convolutional Neural Networks (CNN)
Mehmet Çağrı Aksoy
 
Few shot learning/ one shot learning/ machine learning
Few shot learning/ one shot learning/ machine learningFew shot learning/ one shot learning/ machine learning
Few shot learning/ one shot learning/ machine learning
ﺁﺻﻒ ﻋﻠﯽ ﻣﯿﺮ
 
Introduction to Few shot learning
Introduction to Few shot learningIntroduction to Few shot learning
Introduction to Few shot learning
Ridge-i, Inc.
 
CNN Tutorial
CNN TutorialCNN Tutorial
CNN Tutorial
Sungjoon Choi
 
A Brief History of Object Detection / Tommi Kerola
A Brief History of Object Detection / Tommi KerolaA Brief History of Object Detection / Tommi Kerola
A Brief History of Object Detection / Tommi Kerola
Preferred Networks
 
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
Edge AI and Vision Alliance
 
Introduction to CNN
Introduction to CNNIntroduction to CNN
Introduction to CNN
Shuai Zhang
 
Densely Connected Convolutional Networks
Densely Connected Convolutional NetworksDensely Connected Convolutional Networks
Densely Connected Convolutional Networks
Hosein Mohebbi
 
Regularization
RegularizationRegularization
Regularization
Darren Yow-Bang Wang
 
그림 그리는 AI
그림 그리는 AI그림 그리는 AI
그림 그리는 AI
NAVER Engineering
 
Face recognition using PCA
Face recognition using PCAFace recognition using PCA
Face recognition using PCA
Nawin Kumar Sharma
 
Introduction to object detection
Introduction to object detectionIntroduction to object detection
Introduction to object detection
Brodmann17
 
Mask R-CNN
Mask R-CNNMask R-CNN
Mask R-CNN
Chanuk Lim
 
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Universitat Politècnica de Catalunya
 
Object detection with deep learning
Object detection with deep learningObject detection with deep learning
Object detection with deep learning
Sushant Shrivastava
 
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Simplilearn
 
Image segmentation with deep learning
Image segmentation with deep learningImage segmentation with deep learning
Image segmentation with deep learning
Antonio Rueda-Toicen
 
오토인코더의 모든 것
오토인코더의 모든 것오토인코더의 모든 것
오토인코더의 모든 것
NAVER Engineering
 
Cnn
CnnCnn
Cnn
Nirthika Rajendran
 
eam2
eam2eam2
eam2
butest
 
Heuristic design of experiments w meta gradient search
Heuristic design of experiments w meta gradient searchHeuristic design of experiments w meta gradient search
Heuristic design of experiments w meta gradient search
Greg Makowski
 

Contenu connexe

Tendances

“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
Edge AI and Vision Alliance
 
그림 그리는 AI
그림 그리는 AI그림 그리는 AI
그림 그리는 AI
NAVER Engineering
 
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Universitat Politècnica de Catalunya
 
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Simplilearn
 
오토인코더의 모든 것
오토인코더의 모든 것오토인코더의 모든 것
오토인코더의 모든 것
NAVER Engineering
 

Tendances (20)

A brief introduction to recent segmentation methods
A brief introduction to recent segmentation methodsA brief introduction to recent segmentation methods
A brief introduction to recent segmentation methods
 
Breast Cancer Detection with Convolutional Neural Networks (CNN)
Breast Cancer Detection with Convolutional Neural Networks (CNN)Breast Cancer Detection with Convolutional Neural Networks (CNN)
Breast Cancer Detection with Convolutional Neural Networks (CNN)
 
Few shot learning/ one shot learning/ machine learning
Few shot learning/ one shot learning/ machine learningFew shot learning/ one shot learning/ machine learning
Few shot learning/ one shot learning/ machine learning
 
Introduction to Few shot learning
Introduction to Few shot learningIntroduction to Few shot learning
Introduction to Few shot learning
 
CNN Tutorial
CNN TutorialCNN Tutorial
CNN Tutorial
 
A Brief History of Object Detection / Tommi Kerola
A Brief History of Object Detection / Tommi KerolaA Brief History of Object Detection / Tommi Kerola
A Brief History of Object Detection / Tommi Kerola
 
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
“An Introduction to Data Augmentation Techniques in ML Frameworks,” a Present...
 
Introduction to CNN
Introduction to CNNIntroduction to CNN
Introduction to CNN
 
Densely Connected Convolutional Networks
Densely Connected Convolutional NetworksDensely Connected Convolutional Networks
Densely Connected Convolutional Networks
 
Regularization
RegularizationRegularization
Regularization
 
그림 그리는 AI
그림 그리는 AI그림 그리는 AI
그림 그리는 AI
 
Face recognition using PCA
Face recognition using PCAFace recognition using PCA
Face recognition using PCA
 
Introduction to object detection
Introduction to object detectionIntroduction to object detection
Introduction to object detection
 
Mask R-CNN
Mask R-CNNMask R-CNN
Mask R-CNN
 
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
Image classification on Imagenet (D1L4 2017 UPC Deep Learning for Computer Vi...
 
Object detection with deep learning
Object detection with deep learningObject detection with deep learning
Object detection with deep learning
 
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
Convolutional Neural Network - CNN | How CNN Works | Deep Learning Course | S...
 
Image segmentation with deep learning
Image segmentation with deep learningImage segmentation with deep learning
Image segmentation with deep learning
 
오토인코더의 모든 것
오토인코더의 모든 것오토인코더의 모든 것
오토인코더의 모든 것
 
Cnn
CnnCnn
Cnn
 

Similaire à Face Recognition: From Scratch To Hatch

Face Identification for Humanoid Robot
Face Identification for Humanoid RobotFace Identification for Humanoid Robot
Face Identification for Humanoid Robot
thomaswangxin
 
Learn to Build an App to Find Similar Images using Deep Learning- Piotr Teterwak
Learn to Build an App to Find Similar Images using Deep Learning- Piotr TeterwakLearn to Build an App to Find Similar Images using Deep Learning- Piotr Teterwak
Learn to Build an App to Find Similar Images using Deep Learning- Piotr Teterwak
PyData
 
Intro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft VenturesIntro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft Ventures
microsoftventures
 

Similaire à Face Recognition: From Scratch To Hatch (20)

eam2
eam2eam2
eam2
 
Heuristic design of experiments w meta gradient search
Heuristic design of experiments w meta gradient searchHeuristic design of experiments w meta gradient search
Heuristic design of experiments w meta gradient search
 
backpropagation in neural networks
backpropagation in neural networksbackpropagation in neural networks
backpropagation in neural networks
 
Deep-learning-for-computer-vision-applications-using-matlab.pdf
Deep-learning-for-computer-vision-applications-using-matlab.pdfDeep-learning-for-computer-vision-applications-using-matlab.pdf
Deep-learning-for-computer-vision-applications-using-matlab.pdf
 
Day 4
Day 4Day 4
Day 4
 
Deep Learning in Computer Vision
Deep Learning in Computer VisionDeep Learning in Computer Vision
Deep Learning in Computer Vision
 
B4UConference_machine learning_deeplearning
B4UConference_machine learning_deeplearningB4UConference_machine learning_deeplearning
B4UConference_machine learning_deeplearning
 
Deep learning with TensorFlow
Deep learning with TensorFlowDeep learning with TensorFlow
Deep learning with TensorFlow
 
Troubleshooting Deep Neural Networks - Full Stack Deep Learning
Troubleshooting Deep Neural Networks - Full Stack Deep LearningTroubleshooting Deep Neural Networks - Full Stack Deep Learning
Troubleshooting Deep Neural Networks - Full Stack Deep Learning
 
An Introduction to Deep Learning
An Introduction to Deep LearningAn Introduction to Deep Learning
An Introduction to Deep Learning
 
Introduction to Deep Learning
Introduction to Deep LearningIntroduction to Deep Learning
Introduction to Deep Learning
 
Face Identification for Humanoid Robot
Face Identification for Humanoid RobotFace Identification for Humanoid Robot
Face Identification for Humanoid Robot
 
ML Paper Tutorial - Video Face Manipulation Detection Through Ensemble of CNN...
ML Paper Tutorial - Video Face Manipulation Detection Through Ensemble of CNN...ML Paper Tutorial - Video Face Manipulation Detection Through Ensemble of CNN...
ML Paper Tutorial - Video Face Manipulation Detection Through Ensemble of CNN...
 
Learn to Build an App to Find Similar Images using Deep Learning- Piotr Teterwak
Learn to Build an App to Find Similar Images using Deep Learning- Piotr TeterwakLearn to Build an App to Find Similar Images using Deep Learning- Piotr Teterwak
Learn to Build an App to Find Similar Images using Deep Learning- Piotr Teterwak
 
DeepLearningLecture.pptx
DeepLearningLecture.pptxDeepLearningLecture.pptx
DeepLearningLecture.pptx
 
Machine Learning Essentials Demystified part2 | Big Data Demystified
Machine Learning Essentials Demystified part2 | Big Data DemystifiedMachine Learning Essentials Demystified part2 | Big Data Demystified
Machine Learning Essentials Demystified part2 | Big Data Demystified
 
深度學習在AOI的應用
深度學習在AOI的應用深度學習在AOI的應用
深度學習在AOI的應用
 
Kaggle review Planet: Understanding the Amazon from Space
Kaggle review Planet: Understanding the Amazon from SpaceKaggle review Planet: Understanding the Amazon from Space
Kaggle review Planet: Understanding the Amazon from Space
 
Intro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft VenturesIntro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft Ventures
 
Introduction to computer vision
Introduction to computer visionIntroduction to computer vision
Introduction to computer vision
 

Plus de Eduard Tyantov

Эксплуатация ML в Почте Mail.ru
Эксплуатация ML в Почте Mail.ruЭксплуатация ML в Почте Mail.ru
Эксплуатация ML в Почте Mail.ru
Eduard Tyantov
 
Опыт моделеварения от команды ComputerVision Mail.ru
Опыт моделеварения от команды ComputerVision Mail.ruОпыт моделеварения от команды ComputerVision Mail.ru
Опыт моделеварения от команды ComputerVision Mail.ru
Eduard Tyantov
 
Саморазвитие: как я не усидел на двух стульях и нашел третий
Саморазвитие: как я не усидел на двух стульях и нашел третийСаморазвитие: как я не усидел на двух стульях и нашел третий
Саморазвитие: как я не усидел на двух стульях и нашел третий
Eduard Tyantov
 
Project Management 2.0: AI Transformation
Project Management 2.0: AI TransformationProject Management 2.0: AI Transformation
Project Management 2.0: AI Transformation
Eduard Tyantov
 

Plus de Eduard Tyantov (8)

Эксплуатация ML в Почте Mail.ru
Эксплуатация ML в Почте Mail.ruЭксплуатация ML в Почте Mail.ru
Эксплуатация ML в Почте Mail.ru
 
Опыт моделеварения от команды ComputerVision Mail.ru
Опыт моделеварения от команды ComputerVision Mail.ruОпыт моделеварения от команды ComputerVision Mail.ru
Опыт моделеварения от команды ComputerVision Mail.ru
 
Саморазвитие: как я не усидел на двух стульях и нашел третий
Саморазвитие: как я не усидел на двух стульях и нашел третийСаморазвитие: как я не усидел на двух стульях и нашел третий
Саморазвитие: как я не усидел на двух стульях и нашел третий
 
Project Management 2.0: AI Transformation
Project Management 2.0: AI TransformationProject Management 2.0: AI Transformation
Project Management 2.0: AI Transformation
 
Kaggle Google Landmark recognition
Kaggle Google Landmark recognitionKaggle Google Landmark recognition
Kaggle Google Landmark recognition
 
Deep Learning: Advances Of The Last Year
Deep Learning: Advances Of The Last Year Deep Learning: Advances Of The Last Year
Deep Learning: Advances Of The Last Year
 
Ultrasound nerve segmentation, kaggle review
Ultrasound nerve segmentation, kaggle reviewUltrasound nerve segmentation, kaggle review
Ultrasound nerve segmentation, kaggle review
 
Artisto App, Highload 2016
Artisto App, Highload 2016Artisto App, Highload 2016
Artisto App, Highload 2016
 

Dernier

Abu Dhabi Escorts Service 0508644382 Escorts in Abu Dhabi
Abu Dhabi Escorts Service 0508644382 Escorts in Abu DhabiAbu Dhabi Escorts Service 0508644382 Escorts in Abu Dhabi
Abu Dhabi Escorts Service 0508644382 Escorts in Abu Dhabi
Monica Sydney
 
Indian Escort in Abu DHabi 0508644382 Abu Dhabi Escorts
Indian Escort in Abu DHabi 0508644382 Abu Dhabi EscortsIndian Escort in Abu DHabi 0508644382 Abu Dhabi Escorts
Indian Escort in Abu DHabi 0508644382 Abu Dhabi Escorts
Monica Sydney
 
Call girls Service in Ajman 0505086370 Ajman call girls
Call girls Service in Ajman 0505086370 Ajman call girlsCall girls Service in Ajman 0505086370 Ajman call girls
Call girls Service in Ajman 0505086370 Ajman call girls
Monica Sydney
 
Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...
Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...
Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...
kumargunjan9515
 
一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制
一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制
一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制
pxcywzqs
 
Mira Road Housewife Call Girls 07506202331, Nalasopara Call Girls
Mira Road Housewife Call Girls 07506202331, Nalasopara Call GirlsMira Road Housewife Call Girls 07506202331, Nalasopara Call Girls
Mira Road Housewife Call Girls 07506202331, Nalasopara Call Girls
Priya Reddy
 
一比一原版奥兹学院毕业证如何办理
一比一原版奥兹学院毕业证如何办理一比一原版奥兹学院毕业证如何办理
一比一原版奥兹学院毕业证如何办理
F
 
一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样
一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样
一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样
ayvbos
 
2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs
2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs
2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs
EleniIlkou
 
原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查
原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查
原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查
ydyuyu
 
在线制作约克大学毕业证（yu毕业证）在读证明认证可查
在线制作约克大学毕业证（yu毕业证）在读证明认证可查在线制作约克大学毕业证（yu毕业证）在读证明认证可查
在线制作约克大学毕业证（yu毕业证）在读证明认证可查
ydyuyu
 
call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7
call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7
call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7
9953056974 Low Rate Call Girls In Saket, Delhi NCR
 

Dernier (20)

Abu Dhabi Escorts Service 0508644382 Escorts in Abu Dhabi
Abu Dhabi Escorts Service 0508644382 Escorts in Abu DhabiAbu Dhabi Escorts Service 0508644382 Escorts in Abu Dhabi
Abu Dhabi Escorts Service 0508644382 Escorts in Abu Dhabi
 
Indian Escort in Abu DHabi 0508644382 Abu Dhabi Escorts
Indian Escort in Abu DHabi 0508644382 Abu Dhabi EscortsIndian Escort in Abu DHabi 0508644382 Abu Dhabi Escorts
Indian Escort in Abu DHabi 0508644382 Abu Dhabi Escorts
 
Call girls Service in Ajman 0505086370 Ajman call girls
Call girls Service in Ajman 0505086370 Ajman call girlsCall girls Service in Ajman 0505086370 Ajman call girls
Call girls Service in Ajman 0505086370 Ajman call girls
 
Ballia Escorts Service Girl ^ 9332606886, WhatsApp Anytime Ballia
Ballia Escorts Service Girl ^ 9332606886, WhatsApp Anytime BalliaBallia Escorts Service Girl ^ 9332606886, WhatsApp Anytime Ballia
Ballia Escorts Service Girl ^ 9332606886, WhatsApp Anytime Ballia
 
Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...
Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...
Local Call Girls in Seoni 9332606886 HOT & SEXY Models beautiful and charmin...
 
APNIC Updates presented by Paul Wilson at ARIN 53
APNIC Updates presented by Paul Wilson at ARIN 53APNIC Updates presented by Paul Wilson at ARIN 53
APNIC Updates presented by Paul Wilson at ARIN 53
 
APNIC Policy Roundup, presented by Sunny Chendi at the 5th ICANN APAC-TWNIC E...
APNIC Policy Roundup, presented by Sunny Chendi at the 5th ICANN APAC-TWNIC E...APNIC Policy Roundup, presented by Sunny Chendi at the 5th ICANN APAC-TWNIC E...
APNIC Policy Roundup, presented by Sunny Chendi at the 5th ICANN APAC-TWNIC E...
 
一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制
一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制
一比一原版(Offer)康考迪亚大学毕业证学位证靠谱定制
 
Real Men Wear Diapers T Shirts sweatshirt
Real Men Wear Diapers T Shirts sweatshirtReal Men Wear Diapers T Shirts sweatshirt
Real Men Wear Diapers T Shirts sweatshirt
 
20240509 QFM015 Engineering Leadership Reading List April 2024.pdf
20240509 QFM015 Engineering Leadership Reading List April 2024.pdf20240509 QFM015 Engineering Leadership Reading List April 2024.pdf
20240509 QFM015 Engineering Leadership Reading List April 2024.pdf
 
Mira Road Housewife Call Girls 07506202331, Nalasopara Call Girls
Mira Road Housewife Call Girls 07506202331, Nalasopara Call GirlsMira Road Housewife Call Girls 07506202331, Nalasopara Call Girls
Mira Road Housewife Call Girls 07506202331, Nalasopara Call Girls
 
一比一原版奥兹学院毕业证如何办理
一比一原版奥兹学院毕业证如何办理一比一原版奥兹学院毕业证如何办理
一比一原版奥兹学院毕业证如何办理
 
一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样
一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样
一比一原版(Flinders毕业证书)弗林德斯大学毕业证原件一模一样
 
Tadepalligudem Escorts Service Girl ^ 9332606886, WhatsApp Anytime Tadepallig...
Tadepalligudem Escorts Service Girl ^ 9332606886, WhatsApp Anytime Tadepallig...Tadepalligudem Escorts Service Girl ^ 9332606886, WhatsApp Anytime Tadepallig...
Tadepalligudem Escorts Service Girl ^ 9332606886, WhatsApp Anytime Tadepallig...
 
Trump Diapers Over Dems t shirts Sweatshirt
Trump Diapers Over Dems t shirts SweatshirtTrump Diapers Over Dems t shirts Sweatshirt
Trump Diapers Over Dems t shirts Sweatshirt
 
2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs
2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs
2nd Solid Symposium: Solid Pods vs Personal Knowledge Graphs
 
Nagercoil Escorts Service Girl ^ 9332606886, WhatsApp Anytime Nagercoil
Nagercoil Escorts Service Girl ^ 9332606886, WhatsApp Anytime NagercoilNagercoil Escorts Service Girl ^ 9332606886, WhatsApp Anytime Nagercoil
Nagercoil Escorts Service Girl ^ 9332606886, WhatsApp Anytime Nagercoil
 
原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查
原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查
原版制作美国爱荷华大学毕业证（iowa毕业证书）学位证网上存档可查
 
在线制作约克大学毕业证（yu毕业证）在读证明认证可查
在线制作约克大学毕业证（yu毕业证）在读证明认证可查在线制作约克大学毕业证（yu毕业证）在读证明认证可查
在线制作约克大学毕业证（yu毕业证）在读证明认证可查
 
call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7
call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7
call girls in Anand Vihar (delhi) call me [🔝9953056974🔝] escort service 24X7
 

Face Recognition: From Scratch To Hatch

Notes de l'éditeur

  1. «Detect all faces on the image at all scales and rotations»
  2. http://vis-www.cs.umass.edu/fddb/
  3. http://vis-www.cs.umass.edu/fddb/
  4. https://docs.opencv.org/master/d7/d8b/tutorial_py_face_detection.html , http://blog.dlib.net/2014/02/dlib-186-released-make-your-own-object.html
  5. R-FCN: https://arxiv.org/pdf/1605.06409.pdf , Faster: https://arxiv.org/pdf/1506.01497v3.pdf
  6. «Another approach – HOG features (dlib)» «not far from state-of-the-art (Google reports 94%)»
  7. Offtop: обучается первая сеть на 12x12, вторая – 24x24, 3 – 36x36 – hard negative mining для 2 и 3 для исправления ошибок выдает первая 5x1x1, а на пирамиде тензор, дальше составляется батч для 2 и 3 (батч над батчами не делали) https://kpzhang93.github.io/MTCNN_face_detection_alignment/paper/spl.pdf
  8. «Open source Pre-trained model (Caffe) No original training code «
  9. «but not Torch/PyTorch We’ve retrained the CNN because of PreLu absense» basic- actively devolping
  10. TODO: возможно показать что мы не подряд засовываем в сеть, а сразу в кучу, хз не перегруз ли
  11. табличка в одном месте
  12. https://www.microsoft.com/en-us/research/project/ms-celeb-1m-challenge-recognizing-one-million-celebrities-real-world/
  13. мб добавить
  14. Top results ~ 80% Best – 91.7%
  15. «All faces of person are project to a single point»
  16. Chad Smith
  17. TODO: про альфа лишнее ?
  18. https://ydwen.github.io/papers/WenECCV16.pdf, 65% из статьи
  19. TODO: убрать текст ?
  20. (Better than dlib)
  21. TODO: не выпилить но доделать нормально фотка чувака точки + очки
  22. «пол например не является discriminative – пробовали» 71.5% -> 75.5% или +2 71.5 vs 73, 75,5 из-за датасета
  23. СЛОВА «сказать что норма – это деление на сумму»
  24. To discriminate better -> tighten clusters
  25. http://proceedings.mlr.press/v48/liud16.pdf
  26. http://proceedings.mlr.press/v48/liud16.pdf
  27. “Рассказать про идею созданую по мотивам корпоратиа”
  28. VK – easy API OK – hard access to API, rate-limits Facebook – just closed to mining
  29. Cleaning algorithm Face Detection Face Recognition to get embeddings Hierarchical clustering algorithm with Soften the distance threshold Pick the largest cluster as a person iterate after each model improvement
  30. this leads to improved perf on megaface = 75%?
  31. цифра лучше ? чуть меньше текста или на 2 слайда разбить TODO: мы обнаружили что дети компанкто расположены
  32. https://arxiv.org/pdf/1609.04309.pdf - paper: “Efficient softmax approximation for GPUs”
  33. «какой профит помимо решения проблемы»
  34. https://docs.opencv.org/3.0-beta/doc/py_tutorials/py_imgproc/py_gradients/py_gradients.html It’s a face, but without enough facial information
  35. “If the detector makes a mistake, recognizer gone crazy. Detection errors often form a cluster.”
  36. «софтмакс стремится увеличить скалярное произведение»
  37. maximizing scalar product <w_i, x> for i class We normalize embeddings only for distance
  38. “Social networks contains avatars over users’ lives”
  39. Data we know How many likes Who likes
  40. http://people.cs.uchicago.edu/~larsson/fractalnet/overview.png
  41. block 6 нарисовать
  42. https://arxiv.org/pdf/1611.00822.pdf