Table 7 Computational complexity comparison
From: A deblocking algorithm based on color psychology for display quality enhancement
| Â | Â | ADD | Shift | Compare | Div |
|---|---|---|---|---|---|
Tai’s 4 × 4 filter | Mode decision | 12n | 3n | 14n | – |
|  | Smooth mode | 7n | 12n | – | – |
|  | Intermediate mode | 6n | 4n | 2n | – |
|  | Complex mode | 6n | 6n | 2n | – |
|  | Steep mode | 3n | 4n | – | – |
|  | Corner mode | 24n | 22n | 8n | – |
Chen’s filter | Block size decision stage | 26n | 0.75n | 0.26n | 5n |
|  | No filter mode | – | – | 2n | – |
|  | Low activity mode (strong filter) | 31n | 20n | 2n | – |
|  | Low activity mode (normal filter) | 16n | 10n | – | – |
|  | Intermediate mode | 9n | 10n | – | – |
|  | High activity mode | 6n | 6n | – | – |
|  | Chrominance filter | 10n | 4n | – | – |
Loop filter | Strength measure | 12n | 8.25n | 20n | – |
|  | Filter type decision | 6n | – | 6n | – |
|  | Strong filter | 30n | 20n | – | – |
|  | Standard filter | 11n | 4n | 4n | – |
|  | Chrominance filter | 12n | 12n | – | – |
Proposed filter | Detection-Sobel | 16n | – | 2n | – |
| Â | Detection-wavelate | 8n | 8n | 2n | 2n |
|  | Complex mode | 9n | 24n | – | – |
|  | R mode/G mode/B mode | 9–11n | 10–14n | 3n | – |
|  | Y filter | n | n | n | – |