This device is not compatible.


Deep Learning with Keras for Landscape Recognition

In this project, we’ll learn how to train an image classification model using Keras. We will select the correct model, create the data loaders, configure the hyperparameters, and train, evaluate, and test the model on our image data.

Deep Learning with Keras for Landscape Recognition

You will learn to:

Visualize and examine the image data.

Create the data loaders for training and validation.

Select the correct DL model for the task.

Define the training configuration (loss function, optimizer, etc.).

Train the model and evaluate the performance.

Use the trained model to classify new images.


Deep Learning

Computer Vision

Data Visualization

Convolutional Neural Network


Basic programming skills in Python

Intermediate working knowledge of Keras

Intermediate knowledge of deep learning theory

Basic knowledge of Matplotlib and NumPy


keras logo





Project Description

Deep learning is the most prominent AI technology of the last few years, with outstanding results in computer vision and natural language processing. Keras is among the most popular and user-friendly frameworks for training deep learning models used in academia, industry, and by enthusiasts worldwide.

In this project, we will use Keras on a computer vision task of image classification. We will use landscape images from three different categories (city, mountain, beach) and perform all the steps of the training process. First, we will investigate/understand our data. Next, we will create data loaders for training, validating, and testing the model. Then, we will select an appropriate vision model, a loss function, and an optimizer. Moreover, we will train the model on our data and monitor the performance. Finally, we will use our fine-tuned model to classify new images.

Apart from Keras, we will additionally use NumPy for data manipulation and Matplotlib for visualization.

Project Tasks



Task 0: Get Started

Task 1: Import the Packages


Data Analysis

Task 2: Explore the Structure of the Directories

Task 3: Visualize the Image Data


Data Loaders

Task 4: Create a Data Loader

Task 5: Inspect the Dataset Output

Task 6: Create the Training and Validation Datasets


Create and Compile the Model

Task 7: Create the Model

Task 8: Define the Loss Function and the Metrics

Task 9: Compile the Model

Task 10: Train the Model

Task 11: Evaluate the Model

Task 12: Use the Model on an Unseen Image