Solving Real World Data Science Problems With Python - Computer Vision Edition

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!

Grab it

Dive into a comprehensive tutorial on solving real-world computer vision problems using Python and convolutional neural networks (CNNs). Learn to create and improve models for flower classification, specifically distinguishing "La Eterna" from other types. Explore Tensorflow/Keras libraries to build simple and advanced CNN architectures, implement data augmentation and preprocessing techniques, and utilize Keras Tuner for optimizing network structures. Gain insights on the importance of precision and recall versus accuracy in model evaluation. Follow along with provided code examples, dataset exploration, and step-by-step explanations of CNN concepts and implementation strategies.

Syllabus

- Intro
- Video overview what we’ll be working on
- Code setup GitHub repo & HP challenge link
- Exploring the dataset that we’ll be using
- Reviewing template code starter-code.ipynb
- Installing necessary Python libraries opencv-python, tensorflow
- Reviewing template code part 2
- How we load in the dataset ImageDataGenerator, flow_from_directory
- Building our first classifier convolutional neural net - CNN
- Methods to improve neural network performance MaxPooling, dropout, network architecture
- Quick discussion about importance of precision & recall versus accuracy
- Data augmentation & preprocessing another way to improve performance
- Programmatically finding the best neural network architectures Keras Tuner
- Video recap & conclusion