Local Stereo Matching Using Adaptive Cross-Region-Based Guided Image Filtering with Orthogonal Weights

Table 1

Comparison of bad 2.0 in nonoccluded regions with other local stereo matching algorithms.

Image name	DAWA-F (%)	FASW (%)	IEBIMst (%)	ADSM (%)	IGF (%)	ISM (%)	DoGGuided (%)	Proposed (%)
Adirondack	20.9	26.0	41.3	36.5	36.7	37.3	40.5	17.6
ArtL	20.4	17.2	16.0	20.6	28.0	28.8	23.9	14.9
Jadeplant	44.7	34.3	31.5	35.7	44.6	45.5	42.5	26.1
Motorcycle	16.7	23.0	17.0	27.6	31.5	32.1	30.8	15.5
MotorcycleE	17.1	22.7	20.8	30.5	33.7	34.3	31.6	14.3
Piano	21.6	32.0	30.4	38.8	45.3	44.7	41.8	24.7
PianoL	49.5	43.6	53.9	59.0	52.5	53.1	58.0	42.2
Pipes	21.0	23.7	15.4	26.6	33.5	34.2	28.6	14.5
Playroom	30.5	37.5	42.0	46.8	44.4	44.6	44.6	29.4
Playtable	27.0	43.5	55.0	56.9	59.8	59.8	61.5	36.8
PlaytableP	19.2	28.1	25.6	31.8	44.4	44.5	36.4	19.2
Recycle	19.8	23.2	24.2	29.6	37.8	38.1	33.1	17.9
Shelves	51.3	49.9	52.5	53.3	52.5	51.8	54.6	48.6
Teddy	10.8	10.5	8.57	12.2	21.6	22.0	14.7	8.78
Vintage	39.1	44.1	64.7	52.8	54.9	55.0	57.5	38.9
Australia	47.2	41.7	36.7	40.4	42.7	42.5	45.4	37.5
AustraliaP	13.6	18.1	12.1	20.3	20.1	26.4	23.6	10.8
Bicycle2	13.1	23.1	16.9	27.3	23.7	34.8	30.6	16.3
Classroom2	19.2	27.2	32.5	35.1	32.2	36.1	34.6	17.4
Classroom2E	66.4	40.6	51.0	55.9	45.6	44.5	52.5	44.9
Computer	20.4	19.1	25.3	22.3	28.6	34.4	28.3	17.2
Crusade	30.3	34.9	58.1	56.1	43.0	56.2	59.1	33.5
CrusadeP	33.9	28.1	49.8	50.9	37.2	52.7	53.8	25.2
Djembe	8.73	18.5	11.2	24.2	21.4	25.2	26.4	11.4
DjembeL	48.9	40.8	48.6	58.0	50.9	51.0	60.6	45.4
Hoops	37.8	36.4	56.9	56.3	44.7	52.4	54.7	35.7
Livingroom	26.7	29.3	30.2	36.5	34.7	39.7	38.3	26.6
Newkuba	29.9	28.4	26.8	32.1	31.9	33.3	35.5	23.3
Plants	28.0	31.1	26.9	38.7	37.4	38.8	44.5	23.6
Staircase	36.5	41.0	71.7	69.7	47.1	75.3	72.0	38.4
Weighted average (%)	26.2	28.3	31.5	36.3	36.6	40.2	39.2	23.1