PyImageConf Workshop: Object Detection with Deep Learning

Slide 1

Slide 1 text

Workshop: Object Detection with Deep Learning Understand, implement from scratch & apply.

Slide 2

Slide 2 text

Introduction Who we are 2 Introduction | @tryolabs Agustín Azzinnari Lead Research Engineer @ganitsu Alan Descoins CTO @dekked_

Slide 3

Slide 3 text

Prerequisites for this workshop 3 Introduction Must ● Familiarity with Python. Recommended ● Familiarity with numpy, TensorFlow and Jupyter Notebooks. ● Access to Microsoft-sponsored DSVMs (ssh/Jupyter). Helpful to have ● Basic knowledge of Machine Learning. ● Basics of Deep Learning and Convolutional Neural Networks (CNN).

Slide 4

Slide 4 text

Agenda Introduction 4 Fundamentals: image classification and object detection (Faster R-CNN) Hands on: implement components of Faster R-CNN model Hands on: using Luminoth toolkit for a real world problem

Slide 5

Slide 5 text

Theory 1: Image classification Aka, label this picture for me, please.

Slide 6

Slide 6 text

Image classification What can I do with an image? There’s a cat. 6 (millions of operations later...) (many wasted kWh later...)

Slide 7

Slide 7 text

Image classification From classification to detection Classification There’s a cat in the photo 7 Localization There’s a cat and it’s here Detection There are two cats, here and here

Slide 8

Slide 8 text

What is so hard about this problem? Introducción 8 red green blue 1900 1300 Image classification

Slide 9

Slide 9 text

Challenges of image classification 9 Image classification

Slide 10

Slide 10 text

The Machine Learning approach 10 Image classification Classical models Extract features from the images and use as input to a simple classification algorithm. Deep Learning models Use the images directly as input to a more complex classification algorithm. DATASET

Slide 11

Slide 11 text

Image classification Neural networks and Deep Learning Neural Networks for classification, widely used in the 80s. Convolutional Neural Network (Yann LeCun, 1989) really good for pattern recognition with minimal preprocessing. 11 Handwritten digit recognition LeNet-5, Yann LeCun, 1998.

Slide 12

Slide 12 text

Image classification Convolutional filters A filter that looks at a small region and activates more strongly in the presence of certain pattern. 12 Several filters can detect more complex patterns:

Slide 13

Slide 13 text

Image classification The convolution operation Slide each filter through the image to produce an activation map. 13 Source: https://github.com/vdumoulin/conv_arithmetic Use more filters to detect patterns over activation maps (patterns over patterns over patterns…)

Slide 14

Slide 14 text

Result of a convolutional layer 14 Image classification ... Original image 1900 x 1300 x 3 1900 x 1300 x 1 1900 x 1300 x 64

Slide 15

Slide 15 text

Summarizing information: pooling 15 Image classification ... ... 1900 x 1300 x 64 950 x 850 x 64 2x2 regions become 1x1 (max in each)

Slide 16

Slide 16 text

How (max) pooling looks like 16 Image classification Information corresponding to regions of the image is summarized.

Slide 17

Slide 17 text

Image classification Remaining questions How do we know which filters/patterns to setup? → We learn them. They are regular weights of the network (use backpropagation). How do we know how many filters in each layer? → Hyperparameter of the network (try and see what works best). 17 Source: https://cs231n.github.io/understanding-cnn/

Slide 18

Slide 18 text

18 Finding interesting filters Image classification Learning combinations of filters that are activated (from the activation map) makes it a lot easier to find complex patterns!

Slide 19

Slide 19 text

Final activation map Source: https://blog.heuritech.com/2016/02/29/a-brief-report-of-the-heuritech-deep-learning-meetup-5/ Visualizing a convolutional network Pre-trained: 19 Image classification

Slide 20

Slide 20 text

Theory 2: From classification to detection Aka, tell me what and where, for all you see.

Slide 21

Slide 21 text

Object detection as supervised learning 21 Object detection person bicycle bird bird bird stick door (Pascal VOC)

Slide 22

Slide 22 text

Applications of object detection 22 Object detection CT scan of a lung cancer patient at the Jingdong Zhongmei private hospital in Yanjiao, China's Hebei Province (AP Photo/Andy Wong) Hsieh et al., “Drone-based Object Counting by Spatially Regularized Regional Proposal Networks”, ICCV 2017. Source: Pinterest

Slide 23

Slide 23 text

No content

Slide 24

Slide 24 text

Type of object detection models 24 Object detection Regression based methods Region proposal based methods Single stage prediction of object classes and bounding boxes. Examples: ● You Only Look Once (YOLO, YOLOv2, YOLOv3) ● Single Shot MultiBox Detector (SSD) Two stages: 1. Generate candidate locations using some algorithm. 2. Adjustment of bounding boxes and classification. Examples: ● R-CNN, Fast R-CNN, Faster R-CNN

Slide 25

Slide 25 text

Faster R-CNN Aka, Deep Learning model and its variants work really well.

Slide 26

Slide 26 text

Background Evolution of methods proposed in previous years: 26 Faster R-CNN 2014 R-CNN - Girshick et al. 2015 Fast R-CNN - Girshick. 2016 Faster R-CNN Ren, Girshick et al. “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks”, CVPR 2016.

Slide 27

Slide 27 text

Faster R-CNN A two stage object detector 1. Propose interesting regions (Region Proposal Network, RPN) Where should we look? 2. Analyze proposals & adjust (Region-based CNN, R-CNN) Is this an object? If so, which class? 27 person bicycle

Slide 28

Slide 28 text

First stage (1): Intuition behind regions 28 Faster R-CNN: base network Combination of activation maps encodes spatial information!

Slide 29

Slide 29 text

First stage (2): Resulting set of regions 29 Faster R-CNN: region proposal Potentially hundreds or thousands! Regions are agnostic to particular object classes. → “There might be something here!”

Slide 30

Slide 30 text

30 Faster R-CNN Second stage (1): Deal with variable sizes Resize all the regions to the same dimensions (through Region of Interest Pooling). ... ...

Slide 31

Slide 31 text

31 Faster R-CNN Second stage (2): Classification and adjustment Classification: what type of object is it (or is it background)? → probability distribution Regression: how should I resize the box to better enclose the object? → do it per class ... ...

Slide 32

Slide 32 text

Faster R-CNN Method working summary ● Get an activation map with a CNN. ● Use it to propose interesting regions worth exploring. Associate an objectness score to them. ● Classify regions. Discard those that are background (ie. keep good scores only) Learn how to further adjust for each class of object. 32

Slide 33

Slide 33 text

33 Overview of the network Faster R-CNN RoIP 3. Region of Interest (RoI) Pooling R-CNN 4. Region-based CNN (R-CNN) RPN 2. Region Proposal Network (RPN) 1. Pre-trained base network

Slide 34

Slide 34 text

Hands-on: play around with a Convolutional Network Visualize inner workings of a ResNet.

Slide 35

Slide 35 text

Hands on 1: setup your environment Environment setup Using Microsoft DSVMs SSH access to your instance: ssh cd ~/notebooks/ git clone https://github.com/tryolabs/object-detection-workshop.git cd object-detection-workshop ./download_checkpoint.sh Access Jupyter Hub and pick notebook on object-detection-workshop folder. https://:8000/user// Using your own laptop See the README: https://github.com/tryolabs/object-detection-workshop 35

Slide 36

Slide 36 text

Theory 3: Region Proposal Network (RPN) First stage of Faster R-CNN

Slide 37

Slide 37 text

What we have so far... 37 Faster R-CNN: base network Image of arbitrary size → feature map. Common architectures: ● VGG (16, 19) ● ResNet (50, 101, 152, ...) ● Inception (V2, V3) ● Xception ● MobileNet ● ... 1/16 spatially, 1024 deep for ResNet 101. Feature map 50 37 600 800 CNN (ResNet) 3 1024

Slide 38

Slide 38 text

Proposing regions (1): initial idea 38 Faster R-CNN: region proposal Idea: 1. Look at spatial position and its vicinity. 2. Predict 2 points (x1, y1), (x2, y2) for each location. Issues: ● Can we make the network predict exact pixel coordinates? ● Image dimensions are variable.

Slide 39

Slide 39 text

Proposing regions (2): a better way 39 Faster R-CNN: region proposal 1. Take a single spatial position. 2. Define fixed-size reference box (anchor). 3. Find “closest” GT box. 4. Predict the “objectness” of the region. 5. Learn how to modify the reference box (in relative terms, ie. “double its width”). 6. Repeat for every spatial position.

Slide 40

Slide 40 text

Anchor boxes 40 Faster R-CNN: region proposal For each spatial position of the feature map, generate k fixed anchors (with same center). Ie. 3 scales, 3 aspect ratios (k=9) But choose what’s best for dataset

Slide 41

Slide 41 text

Anchor centers in original image Anchors reference (9 anchors per position) Visualizing anchor boxes All anchors 41 Faster R-CNN: region proposal

Slide 42

Slide 42 text

42 Region Proposal Network (RPN) Faster R-CNN: region proposal Feature map → rectangular proposals + “objectness” score RPN 3x3 conv (pad 1, 512 output channels) 1x1 conv (2k output channels) 1x1 conv (4k output channels) 2k objectness scores 4k box regression scores

Slide 43

Slide 43 text

All positive anchors IoU > 0.7 Anchors batch positive (green), negative (red) 43 Faster R-CNN How does this learn? RPN anchor targets Need positive (foreground) vs negative (background) anchors. Use Intersection over Union (IoU) with ground truth. Faster R-CNN: region proposal

Slide 44

Slide 44 text

What’s missing 44 Faster R-CNN: region proposal Multi-task loss Filtering of proposals ● Use Non-Maximum Suppression (NMS). ● Keep top in “objectness” only. Classification Standard cross-entropy for 2 classes. Box regression Smooth L1 between difference of coordinates (positive anchors).

Slide 45

Slide 45 text

45 Faster R-CNN: region proposal RPN summary: how does a forward pass look like? 1. Run image through base network to get feature map. 2. Run feature map through RPN convolutional layers (3x3, 1x1 & 1x1) a. Obtain objectness and box regression scores for each anchor type and spatial position. b. Use regression scores to adjust each anchor. 3. Sort proposals by objectness score. 4. Apply NMS to remove redundant proposals. Result Set of proposals with associated objectness scores

Slide 46

Slide 46 text

Hands-on: implementing a Region Proposal Network (RPN) First stage of Faster R-CNN, in actual code.

Slide 47

Slide 47 text

Hands on: RPN implementation ● Read, read, read. Docstrings have crucial implementation details, such as shapes and types. ● Comments have hints to help you. ○ We can help you too, don’t be shy and ask! :D Priorities: 1. Make it work (whatever it takes!). 2. Implement it with vectorized numpy. 3. Implement it in pure TensorFlow. a. Can compile and run in GPU. b. You would have to do this for a real implementation. 47

Slide 48

Slide 48 text

Theory 4: RoI Pooling and R-CNN Second stage of Faster R-CNN

Slide 49

Slide 49 text

49 Faster R-CNN Second stage (1): Using our proposals To learn how to adjust anchors, we looked at a small part of the activation map. To decide better, need to look at all activations corresponding to the regions.

Slide 50

Slide 50 text

50 Faster R-CNN Second stage (2): All proposals made equal Turn arbitrarily sized proposals into fixed size vectors / “squares”. Process is called RoI pooling. Allows us to feed into fully connected layer of NN. 7 7

Slide 51

Slide 51 text

Region-based CNN (R-CNN) 51 Faster R-CNN Fixed-size outputs of RoI Pooling→ Faster R-CNN 7x7x1024 probability distribution (N+1 classes) bounding box regressions (N classes) Flatten FC FC bicycle p=0.96 Softmax

Slide 52

Slide 52 text

52 Faster R-CNN person (0.99) bicycle (0.97)

Slide 53

Slide 53 text

Hands-on: implementing the R-CNN Second stage of Faster R-CNN, in actual code.

Slide 54

Slide 54 text

Building a toolkit

Slide 55

Slide 55 text

What is Luminoth? Building a toolkit Open-source deep learning library/toolkit for computer vision object detection. 55 CLI tools Pre-defined models Cloud integration

Slide 56

Slide 56 text

Objectives Building a toolkit 56 “Out-of-the-box” usage Production ready Open source Readable code Extensible and modular

Slide 57

Slide 57 text

$ pip install luminoth $ lumi predict video.mp4 -k car Found 1 files to predict. Neither checkpoint not config specified, assuming `accurate`. Predicting video.mp4 [#############] 100% fps: 5.9 Simplicity as a goal 57 Building a toolkit

Slide 58

Slide 58 text

Data pipeline Debugging Training Data visualization Evaluation Deployment Beyond the model Distributed 58 Building a toolkit Unit testing Monitoring Model

Slide 59

Slide 59 text

Building a toolkit https://github.com/tryolabs/luminoth Using Luminoth 59 $ pip install luminoth $ lumi --help Usage: lumi [OPTIONS] COMMAND [ARGS]... Options: -h, --help Show this message and exit. Commands: checkpoint Groups of commands to manage checkpoints cloud Groups of commands to train models in the cloud dataset Groups of commands to manage datasets eval Evaluate trained (or training) models predict Obtain a model's predictions server Groups of commands to serve models train Train models

Slide 60

Slide 60 text

$ lumi dataset transform --type pascal --data-dir /data/pascal --output /data/ # Create tfrecords for optimizing data consumption. $ lumi train --config pascal-fasterrcnn.yml # Hours of training... $ tensorboard --logdir jobs/ # On another GPU/Machine/CPU. $ lumi eval --config pascal-fasterrcnn.yml # Checks for new checkpoints and writes logs. # Finally. $ lumi server web --config pascal-fasterrcnn.yml # Looks for checkpoint and loads it into a simple frontend/json API server. Luminoth use cycle 60 Building a toolkit

Slide 61

Slide 61 text

Hands-on: Luminoth Build and train a Deep Learning model for traffic

Slide 62

Slide 62 text

Hands on 2: Luminoth for real world object detection Environment setup Read the content in GitHub https://github.com/tryolabs/object-detection-workshop/tree/master/hands-on-2 62

Slide 63

Slide 63 text

Learn more Learning more ● https://tryolabs.com/blog/ ● (Stanford) CS231n: Convolutional Neural Networks for Visual Recognition http://cs231n.stanford.edu/ ● Deep Learning (Goodfellow, Bengio, Courville) http://www.deeplearningbook.org/ ● https://distill.pub/ 63