Algorithm 2: Pseudocode for low-dimensional clustering based on adjacent grid searching.

Input:$\mathrm{dataset} D^d$, noise cells NC, peripheral cells PC, core cells CC

Output: cluster label, cluster density, cluster number m

1: Initialization;

2: Find the adjacent cells for each nonnull cell in CC according to Equation (12);

3: Rank all cells in descending order of density;

4: $\mathrm{CM}\leftarrow$ $\mathrm{cells} \mathrm{number} \mathrm{of} \mathrm{CC}, m\leftarrow$0;

5: For i = 1 to CM Do

6:    If the ith cell in CC is not handled, Then

7:       $m\leftarrow$ m $+1, \mathit{tempCluster}\leftarrow$$\mathrm{the} i^{\mathrm{th}} \mathrm{cell} \mathrm{in} \mathrm{CC}, j\leftarrow$ 1;

8:       While not all cells in tempCluster are handled, Do

9:          Label the jth cell in tempCluster to be the mth cluster;

10:        Add the adjacent cells in CC of the jth cell to tempCluster;

11:        $j\leftarrow$ j + 1;

12:    end While

13: end If

14: end For

15: $\mathrm{PM}\leftarrow$ cells number of PC;

16: While not all cells in PC are handled, Do

17: For i = 1 to PM Do

18:    If the ith cell in PC is not handled, Then

19:      Find its adjacent cells according to Equation (12);

20:      Select the above adjacent cells which are in CC;

21:      Calculate the distance between the cell and its adjacent cells in CC according to                 Equation (15);

22:      Label the ith cell in PC to be the same cluster with its nearest adjacent cell in CC;

23:    end If

24: end For

25: end While

26: Label the data points according to their cells;

27: Calculate the mean density of the cells of each cluster;