Deep learning for object detection using Tensorflow 2

Course Curriculum

Chapter 1: Object detection as a concept in computer vision

Lecture 1: Introduction and course content

Lecture 2: What is object detection for computer vision?

Lecture 3: Object detection can be for multiple objects in the image

Lecture 4: Why deep learning for object detection?

Lecture 5: The 2 categories of neural networks used for object detection

Lecture 6: High level overview of Faster R CNN

Lecture 7: High level overview of SSD

Lecture 8: High level overview of YOLO v3

Chapter 2: How to choose the right neural network for your object detection task

Lecture 1: Introduction

Lecture 2: Speed and accuracy tradeoffs in object detection (reference article)

Lecture 3: Accuracy VS inference time

Lecture 4: Object detector accuracy with respect to feature extractor accuracy

Lecture 5: Accuracy stratified by object size

Lecture 6: How the image resolution affects your object detection model

Lecture 7: How YOLO compares to SSD and Faster R-CNN

Lecture 8: Summary on how to choose the right model for your object detection task

Chapter 3: Software setup

Lecture 1: Linux installation : How to install tensorflow with GPU support (part 1)

Lecture 2: Linux installation : How to install tensorflow with GPU support (part 2)

Lecture 3: Linux installation : How to install Tensorflow 2 object detection API

Lecture 4: Windows installation : Installing miniconda

Lecture 5: Windows installation : Create virtual environment

Lecture 6: Windows installation : Installing tensorflow 2 object detection API

Lecture 7: Windows installation : Installing tensorflow with GPU support

Chapter 4: Data for object detection

Lecture 1: Data preparation for object detection

Lecture 2: The dataset that you will use to build your object detection model

Lecture 3: Downloading and setting up our annotation tool : labelImg

Lecture 4: Annotating your dataset

Lecture 5: Transforming our xml files into one csv file

Lecture 6: Creating a labelmap file for our dataset

Lecture 7: The tool that we will use to generate tfrecords

Lecture 8: Generating tfrecords

Chapter 5: Training an object detection model on your local machine

Lecture 1: Overview of the steps needed to build an object detector

Lecture 2: Transfer learning

Lecture 3: Downloading the pretrained model and getting its corresponding config file

Lecture 4: Preparing your config file

Lecture 5: Running the training and testing for experimentation

Lecture 6: Running Tensorboard to analyse the development of the loss and average precision

Lecture 7: Settings for training and evaluating a Faster RCNN model on your local machine

Lecture 8: Solution : Training and evaluating an SSD based object detection model

Lecture 9: Running the training for SSD object detector

Lecture 10: Other important settings in your config file : anchor boxes

Lecture 11: Other important settings in your config file : data augmentation

Chapter 6: Training object detection API models using Google Cloud AI Platform

Lecture 1: What is cloud computing and what is AI-Platform? (optional)

Lecture 2: Creating a Google Cloud account and creating your first project

Lecture 3: Downloading Google Cloud SDK

Lecture 4: Creating a google bucket and uploading data to it

Lecture 5: Preparing our config file for training on google cloud

Lecture 6: Running the training using Faster RCNN model

Lecture 7: Running the evaluation during the training

Lecture 8: Analyzing the results after the training of Faster RCNN model is finished

Lecture 9: Possible things to do to improve our model performance

Lecture 10: Downloading the trained model and exporting the frozen model from checkpoints

Lecture 11: Running the frozen model on new examples locally

Lecture 12: Solution : Run a training and evaluation using SSD model

Lecture 13: Analyzing the results after the training of SSD model is finished

Chapter 7: YOLO v3 for object detection

Lecture 1: The new setup for training YOLOv3 model

Lecture 2: Installing the necessary requirements to run the code

Lecture 3: How will the dataset change to account for YOLO neural network

Lecture 4: Preparing the dataset : problem with our current dataset

Lecture 5: Preparing the dataset : getting the data from the original source

Lecture 6: Preparing the dataset : transforming our previous dataset to the right format

Lecture 7: Preparing the dataset : adding classes names

Lecture 8: Preparing the dataset : exploring and changing the config file

Lecture 9: Training YOLO v3 based object detection model

Lecture 10: Analyzing the results of the training

Lecture 11: Evaluating our YOLOv3 trained model using new images

Lecture 12: Computing the mean average precision of our trained model

Lecture 13: Other quantitative results

Lecture 14: Saving our model as a SavedModel format for production deployment

Lecture 15: Using our trained model to make predictions on new images

Lecture 16: Steps needed to run the training on google AI platform

Lecture 17: Bonus materials