Lecture 4b Data augmentation for CNN training

Slides:



Advertisements
Similar presentations
A brief review of non-neural-network approaches to deep learning
Advertisements

Face Recognition: A Convolutional Neural Network Approach
Lecture 5: CNN: Regularization
ImageNet Classification with Deep Convolutional Neural Networks
Complex Networks for Representation and Characterization of Object For CS790g Project Bingdong Li 11/9/2009.
Lecture 1: Images and image filtering
K-means Based Unsupervised Feature Learning for Image Recognition Ling Zheng.
Overview of Back Propagation Algorithm
Face Recognition Using Neural Networks Presented By: Hadis Mohseni Leila Taghavi Atefeh Mirsafian.
Convolutional Neural Networks for Image Processing with Applications in Mobile Robotics By, Sruthi Moola.
Presented by: Kamakhaya Argulewar Guided by: Prof. Shweta V. Jain
Images Course web page: vision.cis.udel.edu/cv March 3, 2003  Lecture 8.
Remote Sensing Image Rectification and Restoration
Machine Vision for Robots
Kumar Srijan ( ) Syed Ahsan( ). Problem Statement To create a Neural Networks based multiclass object classifier which can do rotation,
Hurieh Khalajzadeh Mohammad Mansouri Mohammad Teshnehlab
December 4, 2014Computer Vision Lecture 22: Depth 1 Stereo Vision Comparing the similar triangles PMC l and p l LC l, we get: Similarly, for PNC r and.
Digital Image Processing Definition: Computer-based manipulation and interpretation of digital images.
Handwritten Hindi Numerals Recognition Kritika Singh Akarshan Sarkar Mentor- Prof. Amitabha Mukerjee.
Dr. Engr. Sami ur Rahman Digital Image Processing Lecture 9: Rotation, Scaling, Shear, Affine Transformation.
Deep Convolutional Nets
Transformations LESSON 26POWER UP FPAGE 169. Transformations The new image is read as “A prime, B prime, C prime”
CSC321 Lecture 5 Applying backpropagation to shape recognition Geoffrey Hinton.
CSC321: 2011 Introduction to Neural Networks and Machine Learning Lecture 6: Applying backpropagation to shape recognition Geoffrey Hinton.
ImageNet Classification with Deep Convolutional Neural Networks Presenter: Weicong Chen.
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition arXiv: v4 [cs.CV(CVPR)] 23 Apr 2015 Kaiming He, Xiangyu Zhang, Shaoqing.
Yann LeCun Other Methods and Applications of Deep Learning Yann Le Cun The Courant Institute of Mathematical Sciences New York University
Cancer Metastases Classification in Histological Whole Slide Images
Learning to Compare Image Patches via Convolutional Neural Networks
CS 6501: 3D Reconstruction and Understanding Convolutional Neural Networks Connelly Barnes.
Recognition of biological cells – development
Automatic Lung Cancer Diagnosis from CT Scans (Week 2)
Journal Club M Havaei, et al. Université de Sherbrooke, Canada
Every segment is congruent to its image.
Jure Zbontar, Yann LeCun
Every segment is congruent to its image.
CSCI 5922 Neural Networks and Deep Learning: Convolutional Nets For Image And Speech Processing Mike Mozer Department of Computer Science and Institute.
Intelligent Information System Lab
Translation Rotation reflection Dilation Pre Image Image Rigid Motions
Training Techniques for Deep Neural Networks
Efficient Deep Model for Monocular Road Segmentation
Convolutional Networks
CS 698 | Current Topics in Data Science
R-CNN region By Ilia Iofedov 11/11/2018 BGU, DNN course 2016.
Deep Learning Convoluted Neural Networks Part 2 11/13/
Fully Convolutional Networks for Semantic Segmentation
Computer Vision James Hays
Introduction to Neural Networks
CSC 578 Neural Networks and Deep Learning
RGB-D Image for Scene Recognition by Jiaqi Guo
Convolutional Neural Networks
Dog/Cat Classifier Christina Stiff.
Introduction to Deep Learning with Keras
Very Deep Convolutional Networks for Large-Scale Image Recognition
Smart Robots, Drones, IoT
Basics of Deep Learning No Math Required
Spatial Transformer Networks
Papers 15/08.
Use 3D Convolutional Neural Network to Inspect Solder Ball Defects
Essential Question: What can I add to the words slide, flip and turn to more precisely define the rigid-motion transformations – translation, reflection.
Maths Unit 12 – Transformations
CSCI 5922 Neural Networks and Deep Learning: Convolutional Nets For Image And Speech Processing Mike Mozer Department of Computer Science and Institute.
Face Recognition: A Convolutional Neural Network Approach
Automatic Handwriting Generation
VERY DEEP CONVOLUTIONAL NETWORKS FOR LARGE-SCALE IMAGE RECOGNITION
End-to-End Facial Alignment and Recognition
Week 3 Volodymyr Bobyr.
Example of training and deployment of deep convolutional neural networks. Example of training and deployment of deep convolutional neural networks. During.
CSC 578 Neural Networks and Deep Learning
Maths Unit 10 (F) – Transformations
Presentation transcript:

Lecture 4b Data augmentation for CNN training boris.ginsburg@gmail.com

Data Augmentation Training of large network requires a lot of samples. The easiest and most common method to enlarge the dataset using label-preserving transformations: image translations re-scale (both up and down) before crop horizontal reflections elastic deformation with random interpolations ((bilinear, area, nearest neighbor and cubic, with equal probability) (Simard, 2003) photometric distortion and altering the intensities of the RGB channels in training images (A.G. Howard. Some improvements on deep convolutional neural network based image classification, 2013)

Plankton competition The Plankton competition http://www.datasciencebowl.com/: classify grayscale images of plankton into one of 121 classes. The images from camera were processed by a segmentation algorithm to isolate individual organisms, and then cropped. The size of an organism in the resulting images is proportional to its actual size, and does not depend on the distance to the camera. http://benanne.github.io/2015/03/17/plankton.html

Plankton: preprocessing and data augmentation Rescaling all images to the same size zero mean unit variance (ZMUV) normalization Data augmentation: rotation: random with angle in [0°, 360°] translation: random with shift in [-10 ;10] pixels rescaling: random with scale factor [1/1.6 ; 1.6] flipping Shearing (parallel sliding of lines): random with angle [-20° , 20°] stretching: random with stretch factor [1/1.3 ; 1.3]

Data Augmentation preprocessed image Augmented image

ConvNet with cyclic pooling add rotation to data layer 4 parallel identical nets to process rotated images root-mean-square pooling  before output layer

Pre-defined features Pre-defined features: Image size (pixels) Image moments Texture features These features were fed into 2 dense layers and combined the CNN features just before output ‘softmax’ layer

Test-time data transformations For each individual model Applied a quasi-random set of 70 transformations to input image compute predictions across various augmented versions of the input image averaged predictions