Input-Corrective Drawing Pad
Created an input-corrective drawing tablet for those with disabilities that corrects shaky input and uses machine learning (Tensorflow) to classify handwriting • Carried project through engineering design process including stakeholder requirements, specifications document, engineering design, testing, verification, and result reports • Developed interactive art canvas UI • Corrected shaky user input using a line-smoothing algorithm that retained general shapes while correcting error • Engineered a convolutional neural network model in Tensorflow to achieve an accuracy of 98% when reading handwritten digits (approx. 93% from user input)