You will learn the theory behind different algorithms in computer vision in a concise way.
You will be able to deploy your own applications in Computer Vision
How to collect image data from different resources
Description
This course is focused in the application of Deep Learning for image classification and object detection. This course originally was designed in TensorFlow version 1.X but now all codes were updated with TensorFlow version 2.X, mainly by the use of Google Colaboratory(Colab).
If you dont have an available GPU in your local system or you want to expent in an environment without any previous installation or setup, dont worry you can follow the course smootly because all codes were optimized in Google Colab.
The course starts with a concise review of the main concepts in Deep Learning, because this course focused in the application of Deep Learning in the computer vision field.
The main computer vision tasks covered in this course are image classification and object detection.
After reviewing the deep learning theory you will enter in the study of Convolutional Neural Networks (ConvNets) for image classification studying the following concepts and algorithms:
– Image Fundamentals
– Loading images in TensorFlow
– The building blocks of ConvNets such as:
Convolution Operation,
Filters,
Batch Normalization,
ReLU Function,
DropOut,
Pooling Layers,
Dilation,
Shared Weights,
Image Augmentation, etc
– Different ConvNets architectures such as:
LeNet5,
AlexNet,
VGG-16,
ResNet
Inception.
– Many practical applications using famous datasets such as:
Covid19 on X-Ray images,
CIFAR10,
BCCD,
COCO dataset,
Open Images Dataset V6 through Voxel FiftyOne,
ROBOFLOW,
You will also learn how to work and collect image data through web scraping with
Python and Selenium.
Finally in the Object Detection chapter we will explore the theory and the application using Transfer Learning approach using the lastest state of the art algorithms with practical applications. Some of the content in this Chapter is the following:
– Theoretical background for Selective Search algorith,
– Theoretical background for R-CNN, Fast R-CNN and Faster R-CNN,
– Faster R-CNN application on BCCD dataset for detecting blood cells,
– Theoretical background for Single Shot Detector (SSD),
– Training your customs datasets using different models with TensorFlow Object Detection API
– Object Detection on images, videos and livestreaming,
– YOLOv2 theory and practical application in a custom dataset (R2D2 dataset)
– YOLOv3 practical application in a custom dataset (R2D2 and C3PO dataset)
– YOLOv4 theory and practical application in a custom dataset (R2D2 and C3PO dataset)
Finally you will learn how to construct and train your own dataset through GPU computing running Yolo v2, Yolo v3 and the latest
