Doodle Recognition using ConvLSTM2D: A Machine Learning Approach

Doodle recognition has emerged as a key research area in machine learning, aiming to classify hand-drawn sketches by leveraging spatiotemporal data. ConvLSTM2D models, which integrate convolutional layers with LSTM units, are highly effective in capturing both spatial features and the temporal sequence of strokes. This enables the model to understand how doodles evolve over time, improving recognition accuracy. These advancements not only support applications in education, gaming, and art but also promote accessibility by enabling expressive visual communication, especially for individuals with hearing impairments.