RECOMMENDER-SYSTEMS

In this project, we've suggested a system that creates suggestions for the user based on input. It's called a personalised fashion recommender system. This project aims to use an image of a product given as input by the user to generate recommendations instead of the conventional systems that rely on the user's past purchases and history, as people frequently see something they are interested in and tend to look for products that are similar to that. By extracting a number of characteristics from an individual's clothing pictures, we have developed a fashion recommendation system that can figure out that person's preferred clothing style. In order to analyse the images from the Fashion Product Images Dataset, we use convolutional neural networks.The nearest similar images are then recommended using these attributes as input to a similarity model. We have employed model-based recommendation systems, which entail the creation of a model from a collection. In other words, rather than using the entire dataset each time, we extract some data from the dataset and use it as a "model" to generate suggestions. Pre-trained models (CNN-based design) like VGG, Resnet, and Densenet are used for feature extraction. Transfer learning is used to increase accuracy and speed, while CNN is used to recognise the closer images.

DATASET

https://www.kaggle.com/paramaggarwal/fashion-product-images-smallhttps://www.kaggle.com/paramaggarwal/fashion-product-images-small