✨Feature Engineering

Although Fashionic AI will be trained with standard datasets mentioned above, feature engineering data cleaning techniques will be applied to the dataset to reduce redundancy, mistakes, and unimportant data. This is called Data Preprocessing and is widely used in machine learning.

In AI researches about fashion, various feature engineering techniques are used like Color Histograms, Local Binary Patterns (LBP), Histogram of Oriented Gradients (HOG), Scale-Invariant Feature Transformation (SIFT) [27,28]. In addition to these techniques, Deep Learning Methods are used to learn fine-grained features from a dataset. Deep Learning Models are trained by cleaned but not preprocessed data, and the model learns feature sets that are important to predict the outcome. [29-31].

Fashionic AI aims to use Deep Learning Methods to extract features from datasets. However, other techniques will be applied and compared to measure how accurate the model is and how much further improvement can be made.

In summary, Fashionic AI represents the future of the Fashion World. With the help of the Deep Learning and data gathering environment of Inspirest™, Fashionic AI is one of the most promising Artificial Intelligence on Fashion Styling and Recommendation. Fashionic AI will evolve into a unified AI that combines Fashion and Artificial Intelligence.