This project aims to process and render user input photos in a variety of formats and states to a useable cross stitch pattern by using the preprocessing steps used for machine learning models. The delieverables include a web appication that allows for some user customization including: the centering of the input image, adjustments for the type of fabric desired, its size, and the number of colors in the pattern. The outputs from the web application should be a pdf file with color and black and white cross stitch patterns along with a legend determining the appropriate DMC values for color and the associated symbol for that color and the other standard inclusions for cross stitch patterns such as aida count, stitch count per color, and dimensions for the finished product in inches. This output file should automatically be emailed to the end user for reference.
Some longer term goals include the introduction of a feedback system about the generated cross stitch patterns and their ease of use by the end user to generate labeled data for a future machine learning project to determine if the pattern generated is a "good" pattern. The goal of this project is to notify the end user whether or not the uploaded image will produce a satisfactory cross stitch pattern. In addition, hopefully the model will be able to suggest a method of improvement for the input photo, such as higher contrast or brighter light.