Computer Vision

Introduction, Working with Images, CNN Building Blocks, CNN Architectures, Transfer Learning, Visualizations, Semantic Segmentation and Object Detection, Bounding box regressor, CNNs at work: Siamese Network for Metric Learning

 

Introduction –

·         Courseoutline -Computer Vision.pdf

 

Working with Images

1.       Working with Images_Introduction

2.       Working with Images – Digitization, Sampling, and Quantization

3.       Working with images – Filtering

4.       Hands-on Python Demo: Working with images

5.       Working with Images.ipynb

CNN Building Blocks

1.       Introduction to Convolutions

2.       2D convolutions for Images

3.       Convolution – Forward

4.       Convolution – Backward

5.       Transposed Convolution and Fully Connected Layer as a Convolution

6.       Pooling : Max Pooling and Other pooling options

7.       Hands-on Keras Demo:

1.       MNIST CNN Building Blocks code walk-through

2.       MNIST_CNN_Cloud.ipynb

3.       MNIST_CNN_Colab.ipynb

Project Work

1.       Project Description.pdf

2.       SVHN_CNN_Colab.ipynb

3.       SVHN_CNN_Cloud.ipynb

4.       SVHN_CNN_Colab_Solution.ipynb

 

CNN Architectures, Transfer Learning, Visualizations

1.       CNN Architectures and LeNet Case Study

2.       Case Study : AlexNet

3.       Case Study : ZFNet and VGGNet

4.       Case Study : GoogleNet

5.       Case Study : ResNet

6.       Transfer Learning Principles and Practice

7.       Hands-on Keras Demo: SVHN Transfer learning from MNIST dataset

1.       SVHN_CNN_Transfer.ipynb

2.       cnn_mnist_weights.h5

8.       Visualization (run pacakge, occlusion experiment)

1.       Hands on demo -T-SNE

2.       t-SNE MNIST.ipynb

 

Semantic Segmentation and Object Detection

1.       CNNs at Work – Semantic Segmentation

2.       Semantic Segmentation process

3.       U-Net Architecture for Semantic Segmentation

4.       Hands-on demo – Semantic Segmentation using U-Net

5.       UNet-StyleV1.ipynb

6.       Other variants of Convolutions

7.       Inception and Mobile Net models

8.       CNNs at Work – Object Detection with region proposals

9.       CNNs at Work – Object Detection with Yolo and SSD

10.   Hands-on demo- Bounding box regressor

1.       SingleBoundingBoxV1.ipynb

2.       validation.csv

3.       train.csv

Project Work

1.       DLCP Project 2 Brief.pdf

2.       Files Required for Face Detection

1.       Files_required_for_face_detection.zip

3.       FACE_DETECTION_Questions_final.ipynb

4.       FACE_DETECTION_Solution.ipynb

CNNs at work: Siamese Network for Metric Learning

1.       Metric Learning

2.       Siamese Netwrok as metric learning

3.       How to train a Neural Network in SIamese way

4.       Hands-on demo – Siamese Network

1.       Few Shot Learning – V1.ipynb

2.       images_background.zip

3.       images_background_small1.zip

4.       images_background_small2.zip

5.       images_evaluation.zip

6.       train.pickle

7.       val.pickle

Project Work

1.       DLCP Project 2 Brief.pdf

2.       Files Required for Face recognition

1.       new_FACE_RECOGNITION_Questions_updated.ipynb

2.       new_FACE_RECOGNITION_Solution-1.ipynb

3.       Siamese_SigNet_BHSig260.ipynb

4.       Object_detection_fit.ipynb