Do you know the feeling of not having anything nice to wear? You want a new outfit but you have absolutely no idea what to buy? You want to try a new style, but you have no inspiration? The Magic Mirror developed at ScaDS.AI Dresden/Leipzig works with various algorithms based on deep learning, such as image classification, speech recognition, natural language processing and image generation to provide clothing recommendations and improve shopping experiences. It becomes clear here that an Artificial Intelligence (AI) can only be trained on one specific aspect, which is why multiple algorithms are implemented in this demonstrator.

Using the Magic Mirror

To use the Magic Mirror, users must stand in front of a large display, which has an integrated camera. By pressing the shutter button, the camera takes a picture of the user and segments the pieces of clothing the user is wearing. By selecting one of them, the AI can classify the piece. After the clothing pieces are segmented, the user will need to speak into a microphone. The user is asked to describe what he or she would like to wear. With speech recognition, the AI will generate items that match the user’s description and dress them accordingly.

In the future, the Magic Mirror could be developed as an app for individual use. Applications like this could also be installed in shops as touchpoints. We are currently working on improving the Magic Mirror application. We will keep you updated on further progress.