Black hole and strange emission glitches when using oil paint post process effect

[Miszla]

I’ve implemented an Oil Paint Post Process effect from this article and I run into some issues.

1. The sky behaves… weird
   

   

2. The emission materials also look strange
   

Here is a post process code:

int TexIndex = 14;

 int intensityCount[10];
 float avgR[10];
 float avgG[10];
 float avgB[10];


 for (int iLevel = 0; iLevel < 10; iLevel++){
     intensityCount[iLevel] = 0;
     avgR[iLevel] = 0.0;
     avgG[iLevel] = 0.0;
     avgB[iLevel] = 0.0;
  }


 //COUNT INTENSITIES
 uv *= 0.5;
 for (int i = 0; i < radius; ++i)
 {
int offsetI = -1 *(radius / 2) + i;
float v = uv.y + offsetI * invSize.y;
int temp = i * radius;
     for (int j = 0; j < radius; ++j)
    {
int offsetJ = -(radius / 2) + j;
float u = uv.x + offsetJ * invSize.x;
float2 uvShifted = uv + float2(u, v);
float3 tex = SceneTextureLookup(uvShifted, TexIndex, false);

float currentIntensity = ((tex.r + tex.g + tex.b) / 3 * 10);
        intensityCount[currentIntensity]++;
        avgR[currentIntensity] += tex.r;
        avgG[currentIntensity] += tex.g;
        avgB[currentIntensity] += tex.b;
    }
  }


 float maxIntensity = 0;
 int maxIndex = 0;

 //FIND CORRECT MOST COMMON INTENSITY LEVEL
 for(int cLevel = 0; cLevel < 10; cLevel++){
     if(intensityCount[cLevel] > maxIntensity){
         maxIntensity = intensityCount[cLevel];
         maxIndex = cLevel;
     }
  }

 float newR = avgR[maxIndex] / maxIntensity;
 float newG = avgG[maxIndex] / maxIntensity;
 float newB = avgB[maxIndex] / maxIntensity;

 float4 res = float4(newR, newG, newB, 1.0);

 return res;

Do you know what cas possibly cause it and for what should I look for to fix it? I have very limited knowledge about how this code works but with what I know, it shouldn’t behave so strangely