References in periodicals archive ?
1) We propose a new optimization framework RLDB (Robust-LDB).
The pipeline of the proposed RLDB is described in Fig.
In this section, we present the experimental results to evaluate the proposed RLDB descriptor and compare its performance with state-of-the-art descriptors on public datasets.
For RLDB, we randomly chose 50K matching pairs and 200K non-matching pairs from Brown patch datasets [27, 34, 35] to perform the supervised training in Section 3.2, and set parameters [t.sub.c] and [t.sub.e] to 0.45 and 1.2.
We first perform an evaluation using the benchmark from [22, 34] to investigate the discriminative capability of RLDB descriptor.
7 shows, RLDB descriptor outperforms original LDB distinctly.
Acronyms browser ?
Full browser ?
- RLC circuit
- RLE Compression