Imagine that you want to train a deep learning model to predict what is displayed in a specific image. Perhaps you would like to accurately distinguish an image of a t-shirt from an image of a pair of pants, or to distinguish between attributes such as blue and red.
For online retailers, image classification is an essential part of the customer offering and experience, and in turn, the business operations. Being able to automatically tag an image of a product with the correct attribute labels, enables the retailer to organize its product catalog. It can also allow for the customer to filter products according to attributes while in a specific category, in order to more easily find what they are looking for. This is called image classification, a task of which deep learning is widely used to solve.
Although the image classification in this case will identify fashion items, it can be trained on any type of image you require. For example, in healthcare, this technique could use brain scans as the input, and predict if the said brain scan contains a tumor. This technique could also be used to classify a personal photo library, much as Apple or Google do with their photo applications.
Click here for the full text version of the tutorial.
In our tutorial “Classifying images of clothes,” you are guided through the task of solving a typical classification problem using a convolutional neural network (CNN). We use the Fashion MNIST dataset, consisting of small, 28x28-pixel grayscale images of clothing and accessories such as shirts, bags, shoes and other fashion items.
The dataset consists of 70,000 images derived from the online retailer Zalando. Each image is annotated with a label indicating the correct type of fashion item.
What you will learn
/ Building and training a model for solving a typical classification problem, predicting what type of clothing a specific image depicts
/ Using preloaded deep learning model templates, called snippets
/ Exploring the key concepts of the Peltarion Platform
Want to get started with creating your own image classification model? Click here for the full text version of the tutorial.