Skip to main content

Table 2 Performance comparison of three methods on the fountain and HerzJesu dataset

From: Efficient cost aggregation for feature-vector-based wide-baseline stereo matching

Method

Fountain-A

Fountain-B

HerzJesu

Time (sec)

 

Precision

Recall

Precision

Recall

Precision

Recall

 

Min et al. [14]

0.904

0.838

0.831

0.765

0.907

0.845

128

Census transform [41]

0.981

0.788

0.975

0.714

0.978

0.759

105

Proposed method

0.953

0.809

0.934

0.751

0.967

0.823

53