ReacTex: using React Native and Neural Networks to recognize handwritten equations
summary
Builds a handwriting-to-LaTeX pipeline that captures pen strokes in React Native, rasterizes them into a bitmap and feeds them to a TensorFlow neural net with two hidden layers trained as a classifier; strokes get sent to a web React equation editor which inserts rendered LaTeX symbols. Code uses React SVG canvas, Node to glue pieces, and TensorFlow for the model. Claims working end-to-end but gives no accuracy percentage or dataset size.
AI-written summary · about
chapters (5)
related talks
44mBuilding Handwriting Recognition System for the NY Times Crossword
40mBuilding a Handwriting Recognition System
25mBuilding a Handwriting Recognition System for the New York Times Crossword
26mSmart Apps: Machine Learning in React Native
10mMachine Learning And React Native
44mHello Deep Learning
26mBuild your Machine Learning Model on Edge with React Native
24mReact Native Ignite: Building an AI app