public DistanceAndPoint NearestFrom(Vector2 point, DistanceAndPoint nearest)
{
var result = nearest;
// 自分との距離
var dist = (Value - point).sqrMagnitude;
// 現在の最小距離の範囲にある領域が探索候補
if (dist < result.Distance)
{
result = new DistanceAndPoint(dist, Value);
}
// 存在する可能性のある Area を探索
var areaFlags = new bool[] { false, false, false, false };
var index = (int)Detect(new Vector2(point.x + result.Distance, point.y));
areaFlags[index] = true;
index = (int)Detect(new Vector2(point.x - result.Distance, point.y));
areaFlags[index] = true;
index = (int)Detect(new Vector2(point.x, point.y + result.Distance));
areaFlags[index] = true;
index = (int)Detect(new Vector2(point.x, point.y - result.Distance));
areaFlags[index] = true;
for (var i = 0; i < 4; i++)
{
if (!areaFlags[i])
{
continue;
}
var tree = children[i];
if (tree == null)
{
continue;
}
var childResult = tree.NearestFrom(point, result);
var d = childResult.Distance;
if (d >= result.Distance)
{
continue;
}
result = childResult;
}
return(result);
}
private Vector2 BlueNoise()
{
var candidates = Enumerable.Range(0, CandidateCount).Select(_ => WhiteNoiseGenerator.Next()).ToList();
var sample = samples.Last();
var best = new DistanceAndPoint(0, sample);
for (var j = 0; j < CandidateCount; j++)
{
var candidate = candidates[j];
var nearest = quadTree.NearestFrom(candidate, new DistanceAndPoint(int.MaxValue, sample));
if (nearest.Distance > best.Distance)
{
best = new DistanceAndPoint(nearest.Distance, candidate);
}
}
quadTree.Insert(best.Point);
samples.Add(best.Point);
return(best.Point);
}
private Vector2 FilledBlueNoise()
{
if (allSamplesCount == 0)
{
return(Vector2.zero);
}
var candidateIndexes = Enumerable.Range(0, CandidateCount)
.Select(_ => Random.Range(0, allSamplesCount));
var candidates = candidateIndexes.Select(x =>
new
{
index = x,
point = allSamples[x]
}).ToList();
var sample = samples.Last();
var best = new DistanceAndPoint(0, sample);
var bestIndex = 0;
for (var j = 0; j < CandidateCount; j++)
{
var candidate = candidates[j];
var nearest = quadTree.NearestFrom(candidate.point, new DistanceAndPoint(int.MaxValue, sample));
if (nearest.Distance > best.Distance)
{
best = new DistanceAndPoint(nearest.Distance, candidate.point);
bestIndex = candidate.index;
}
}
quadTree.Insert(best.Point);
samples.Add(best.Point);
allSamples[bestIndex] = allSamples[allSamplesCount - 1];
allSamplesCount--;
return(best.Point);
}
