Node splitting algorithms in tree-structured high-dimensional indexes for similarity search