Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

Drawing Recognition with CNNs for Sketches

Go to class Write review

Overview

The course focuses on the application of Convolutional Neural Networks (CNNs) for recognizing hand-drawn sketches. It covers the architecture of CNNs, how they can be trained on sketch datasets, and practical exercises to reinforce the concepts learned.

Syllabus

Unit 1: Building the CNN Model for Sketch Recognition

Adding Dense Layer to the Model
Building the CNN Foundation
Completing the CNN Architecture

Unit 2: Training the CNN Model

Compiling and Training Your Sketch Recognizer
Evaluating Model Performance and Generalization
Saving and Loading Your Sketch Recognizer

Unit 3: Evaluating the Model and Visualizing the Predictions

Visualizing Loss and Accuracy Together
Creating Confusion Matrix for Sketch Recognition
Finding Most Confused Category Pairs
Detailed Metrics for Sketch Recognition Performance
Visualizing Misclassified Sketches in Action

Unit 4: Improving the Model Performance

Preventing Overfitting with Dropout and EarlyStopping
Automatic Training Termination with EarlyStopping
Enhancing CNN Models with Regularization Techniques

Reviews

Start your review of Drawing Recognition with CNNs for Sketches