SpatialSearchEngine

A C++ implementation of a KD Tree supporting efficient k-Nearest Neighbor (KNN) search and classification on the MNIST dataset.

This project demonstrates efficient multidimensional spatial indexing using a balanced KD Tree and branch-and-bound nearest neighbor search.

Features

• Balanced KD Tree construction
• Fast KNN search
• Point insertion and lookup
• MNIST digit classification
• Efficient multidimensional search

Project Structure

• src/Point.h – Point representation in k-dimensional space.
• src/KDTree.h – KD Tree implementation.
• src/BoundedPQueue.h – Priority queue for KNN search.
• src/main.cc – Runs KD Tree operations and KNN classification.

References

• KD Tree - Wikipedia
• MNIST Dataset

Build & Run

make
./main

---

Built using C++ and the STL to demonstrate efficient spatial indexing and nearest-neighbor search algorithms.