Binary Mask Neighbor Majority Filter

Description

Apply a strict 3×3 binary majority filter with optional center inclusion and zero-valued outside neighbors.

A binary grid represents foreground as numeric one and background as zero. A stencil is a set of integer [rowOffset,columnOffset] positions relative to a cell. Connectivity describes allowed neighbor steps: four-neighbor connectivity uses orthogonal steps, while eight-neighbor connectivity also includes diagonals. Finite image borders and asymmetric stencils affect the result.

Apply a strict 3×3 binary majority filter with optional center inclusion and zero-valued outside neighbors.

When to use this tool

Use a strict local vote to remove sparse binary noise or compare center-including and center-excluding neighborhoods.

How to use the tool and how it works

Enter mask, includeCenter using the definitions below.
Run the operation. Resolve input-format errors before interpreting an output.
Read result.mask, using the stated boundary, ordering, sign and absence conventions.

Count ones in the surrounding 3×3 neighborhood, including the center only when includeCenter is true. Set the output to one only if the count exceeds half of the fixed neighborhood size: nine including the center, eight otherwise.^[1]

Input field	Meaning and units	Accepted input
mask	Rectangular numeric 0/1 rows, up to 64×64; row is the first coordinate and column the second.	Dense numeric rows, each dimension at most 128; the stated shape rules also apply.
includeCenter	Whether the current cell participates in the fixed-size majority neighborhood: nine cells if true, eight if false.	Boolean true or false.

Output field	Interpretation
result	Object containing the fields below.
result.mask	Result grid in original frame dimensions, with numeric cell meanings and any signed subtraction convention described above.

json

{
  "input": {
    "mask": [
      [
        0,
        0,
        0
      ],
      [
        0,
        1,
        0
      ],
      [
        0,
        0,
        0
      ]
    ],
    "includeCenter": true
  },
  "output": {
    "result": {
      "mask": [
        [
          0,
          0,
          0
        ],
        [
          0,
          0,
          0
        ],
        [
          0,
          0,
          0
        ]
      ]
    }
  }
}

Limitations and assumptions

Outside-frame neighbors count as zero while the denominator remains nine or eight. Ties become zero, so edge behavior differs from a majority over only existing neighbors.
Masks are rectangular dense arrays of numeric 0/1 cells, with at most 64 rows and 64 columns. Empty frames are allowed where no in-bounds seed is required. Stencils are nonempty lists of distinct integer [rowOffset,columnOffset] pairs, each coordinate between -16 and 16. Outside-frame samples are zero at every stage.
Every declared outer field is required and unknown outer fields are rejected. Strings and JSON text have size limits, lists are bounded to 128 entries, and JSON nesting/node counts are bounded. Tighter operation-specific bounds stated above take precedence.
Numeric outputs must remain finite and within safe-integer magnitude. Original record indexes and cell coordinates are zero-based; stated half-open ranges exclude their end.

Alternative or Complementary methods

Choose a component operation for connected regions and an explicit stencil for local shape tests. Fix connectivity and border policy before comparing masks; convert cell counts to physical area separately when pixel spacing is known.

References

Tool behavior and example: Binary Mask Neighbor Majority Filter — tool.io

Similar or alternative tools

Binary Mask Isolated Cell Remover
Remove foreground components containing exactly one cell under the chosen connectivity.
Binary Mask Explicit Stencil Opening
Apply explicit-stencil erosion followed by dilation within the original finite binary-grid frame.