‘CADSketchNet’ - An Annotated Sketch dataset for 3D CAD Model Retrieval with Deep Neural Networks

Ongoing advancements in the fields of 3D modelling and digital archiving have led to an outburst amount data stored digitally. Consequently, several retrieval systems been developed depending on type these databases. However, unlike text or images, performing a search for models is non-trivial. Among models, retrieving Engineering/CAD mechanical components even more challenging due presence holes, volumetric features, sharp edges etc., which make CAD domain unto itself. The research work presented this paper aims at developing dataset suitable building system based deep learning. from available databases are collected, computer-generated sketch data, termed ‘CADSketchNet’, has prepared. Additionally, hand-drawn sketches also added CADSketchNet. Using images dataset, evaluating performance various engine that accepts image as input query. Many experimental constructed tested These experiments, along with model architecture, choice similarity metrics reported results.