Convoluted fashion classifier prediction using Neural Networks

Abstract

Abstract The conventional way of choosing a fashion catalogue has a lot of drawbacks and is yet to be resolved. Struck in a dilemma of what to choose and what not to, not wearing fashion items that are appropriate, and lured by terms that are misleading such as 50 percent off limited period offer, when in reality the prize was never higher in the first place. These are the flaws that go under the radar for humans, but that is not the case with machines. The proposed system when given an image of a fashion item, by the user makes use of a Convolution Neural Network to find the type of item, then suggest items based on occasion. The system also gives you the prize of that particular item over a period of time in various e commerce sites, and if there is any chance of drop in prize in the future. This eliminates the human part of errors while choosing a fashion catalogue and makes everything automated. The next advantage of our site is to resell our old clothes at best prize so our used clothes are not wasted and it can be used by others.