fix(mask): keep only largest connected slab component

Sheet gaps after black/green removal left stray chunks in the bitset.
After keep predicates, zero all but the largest 4-connected component
and recompute slabPx before bbox / slabBounds.
This commit is contained in:
grokbot_ty
2026-09-26 08:33:19 -07:00
parent b55344ec05
commit 0dd5d255ec
3 changed files with 154 additions and 0 deletions
@@ -0,0 +1,68 @@
import { test } from 'node:test';
import assert from 'node:assert/strict';
import { keepLargestComponent } from './largestComponent.ts';
test('keepLargestComponent: two blobs → only larger remains', () => {
// 10×6 grid: large 4×3 block (12 px) at left, small 2×2 (4 px) at right
const W = 10, H = 6;
const bitset = new Uint8Array(W * H);
for (let y = 1; y <= 3; y++)
for (let x = 1; x <= 4; x++) bitset[y * W + x] = 1;
for (let y = 2; y <= 3; y++)
for (let x = 7; x <= 8; x++) bitset[y * W + x] = 1;
const before = bitset.reduce((a, c) => a + c, 0);
assert.equal(before, 16);
const slabPx = keepLargestComponent(bitset, W, H);
assert.equal(slabPx, 12, 'slabPx is pixel count of largest blob');
// Large blob intact
for (let y = 1; y <= 3; y++)
for (let x = 1; x <= 4; x++)
assert.equal(bitset[y * W + x], 1, `large (${x},${y}) kept`);
// Small blob zeroed
for (let y = 2; y <= 3; y++)
for (let x = 7; x <= 8; x++)
assert.equal(bitset[y * W + x], 0, `small (${x},${y}) removed`);
assert.equal(bitset.reduce((a, c) => a + c, 0), 12);
});
test('keepLargestComponent: empty bitset → 0', () => {
const bitset = new Uint8Array(4 * 4);
assert.equal(keepLargestComponent(bitset, 4, 4), 0);
});
test('keepLargestComponent: single blob unchanged', () => {
const W = 5, H = 5;
const bitset = new Uint8Array(W * H);
for (let y = 1; y <= 3; y++)
for (let x = 1; x <= 3; x++) bitset[y * W + x] = 1;
const slabPx = keepLargestComponent(bitset, W, H);
assert.equal(slabPx, 9);
assert.equal(bitset.reduce((a, c) => a + c, 0), 9);
});
test('keepLargestComponent: 4-connected — diagonal touch is two components', () => {
// Two pixels only touching diagonally; larger is the two-pixel L on the left
// of a separate single pixel that is diagonal to nothing of equal size...
// Layout: three pixels as an L (size 3) and one diagonal fleck (size 1)
// . # .
// # # .
// . . #
const W = 3, H = 3;
const bitset = new Uint8Array(W * H);
bitset[0 * W + 1] = 1;
bitset[1 * W + 0] = 1;
bitset[1 * W + 1] = 1;
bitset[2 * W + 2] = 1; // diagonal-only to the L
const slabPx = keepLargestComponent(bitset, W, H);
assert.equal(slabPx, 3);
assert.equal(bitset[2 * W + 2], 0, 'diagonal fleck removed under 4-conn');
assert.equal(bitset[0 * W + 1], 1);
assert.equal(bitset[1 * W + 0], 1);
assert.equal(bitset[1 * W + 1], 1);
});
+80
View File
@@ -0,0 +1,80 @@
/**
* Keep only the single largest connected component in a slab bitset.
* Sheet gaps can leave stray chunks after sheet removal; those must not
* affect bbox / slabBounds / slabPx.
*
* Connectivity: 4-connected (N/E/S/W). Mutates bitset in place; returns
* remaining slab pixel count.
*/
export function keepLargestComponent(
bitset: Uint8Array,
w: number,
h: number,
): number {
const n = w * h;
if (n === 0) return 0;
// labels[i] = component id (1-based), 0 = background / unvisited
const labels = new Int32Array(n);
const sizes: number[] = [0]; // sizes[id]
let nextId = 0;
const qx = new Int32Array(n);
const qy = new Int32Array(n);
for (let y = 0; y < h; y++) {
const row = y * w;
for (let x = 0; x < w; x++) {
const i = row + x;
if (!bitset[i] || labels[i]) continue;
nextId++;
let size = 0;
let qh = 0, qt = 0;
qx[qt] = x;
qy[qt] = y;
qt++;
labels[i] = nextId;
while (qh < qt) {
const cx = qx[qh];
const cy = qy[qh];
qh++;
size++;
// 4-connected neighbors
const nbrs: [number, number][] = [
[cx - 1, cy],
[cx + 1, cy],
[cx, cy - 1],
[cx, cy + 1],
];
for (const [nx, ny] of nbrs) {
if (nx < 0 || ny < 0 || nx >= w || ny >= h) continue;
const ni = ny * w + nx;
if (!bitset[ni] || labels[ni]) continue;
labels[ni] = nextId;
qx[qt] = nx;
qy[qt] = ny;
qt++;
}
}
sizes[nextId] = size;
}
}
if (nextId === 0) return 0;
let bestId = 1;
let bestSize = sizes[1];
for (let id = 2; id <= nextId; id++) {
if (sizes[id] > bestSize) {
bestSize = sizes[id];
bestId = id;
}
}
if (nextId === 1) return bestSize;
for (let i = 0; i < n; i++) {
if (bitset[i] && labels[i] !== bestId) bitset[i] = 0;
}
return bestSize;
}
+6
View File
@@ -2,6 +2,7 @@ import { minAreaRect, slabBounds } from '../geometry/rect';
import type { MaskParams, WorkerIn, WorkerOut } from '../capture/capture.types';
import { decodeOrientedBitmap } from '../decodeOrientedBitmap';
import { greenKeep, blackKeep, autoPickSheet } from '../mask/sheetKeep';
import { keepLargestComponent } from '../mask/largestComponent';
let bitmap: ImageBitmap | null = null;
let data: Uint8ClampedArray | null = null;
@@ -33,6 +34,11 @@ function runMask(params: MaskParams, id: number): WorkerOut {
if (slabPx === 0) {
return { id, kind: 'ready', slabPx: 0, bitset, bbox: { corners: [], longSidePx: 0, angleDeg: 0 }, slabBounds: { x: 0, y: 0, w: 0, h: 0 } };
}
// Sheet gaps leave stray chunks; lock onto the single largest slab object.
slabPx = keepLargestComponent(bitset, w, h);
if (slabPx === 0) {
return { id, kind: 'ready', slabPx: 0, bitset, bbox: { corners: [], longSidePx: 0, angleDeg: 0 }, slabBounds: { x: 0, y: 0, w: 0, h: 0 } };
}
const bbox = minAreaRect(bitset, w, h);
if (!bbox) {
return { id, kind: 'ready', slabPx: 0, bitset, bbox: { corners: [], longSidePx: 0, angleDeg: 0 }, slabBounds: { x: 0, y: 0, w: 0, h: 0 } };