An optimisation approach to the design of a fast, compact and distinctive binary descriptor

Abstract

The desired local feature descriptor should be distinctive, compact and fast to compute and match. Therefore, many computer vision applications use binary keypoint descriptors instead of floating-point, rich techniques. In this paper, an optimisation approach to the design of a binary descriptor is proposed, in which the detected keypoint is described using several, scale-dependent patches. Each such patch is divided into disjoint blocks of pixels, and then, binary tests between blocks' intensities, as well as their gradients, are used to obtain the binary string. Since the number of image patches and their relative sizes influence the descriptor creation pipeline, a simulated annealing algorithm is used to determine them, optimising recall and precision of keypoint matching. The simulated annealing is also used for dimensionality reduction in long binary strings. The proposed approach is extensively evaluated and compared with SIFT, SURF and BRIEF on public benchmarks. Obtained results show that the binary descriptor created using the resulted pipeline is faster to compute and yields comparable or better performance than the state-of-the-art descriptors under different image transformations.

RLDB: Robust Local Difference Binary Descriptor with Integrated Learning-based Optimization

“…The experimental results show that using this framework leads to significant improvements over original LDB and maintains its advantage in efficiency. Furthermore, LDB is a typical region-based descriptor and a number of binary descriptors [41,42] have similar structure with it, thus the proposed framework is also meaningful or illuminating for the design and usage of other binary descriptors.…”

An approach to vision-based localisation with binary features for partially sighted people

“…Among recently introduced techniques, Binary Online Learned Descriptor (BOLD) [4] is independently optimised for each image patch, and Receptive Fields Descriptor (RFD) [9] thresholds fields' responses of rectangular or Gaussian pooling regions. In Optimised Binary Robust fAst Features (OBRAF) [28], up to 12 image patches with different scaledependent sizes are divided into 3 × 3 pixel blocks and then pairwise tests on intensities and directional gradients are performed. In that solution, the binary string is reduced using a simulated annealing algorithm, or only four patches are used, leaving intensity tests in a simplified version of this descriptor [29].…”

“…Well-performing binary descriptors often require dimensionality reduction [28,39] or learning which can be prone to the overfitting, e.g. BinBoost showed outstanding performance in patch-based benchmarks, while obtaining mediocre results in typical image matching tests [4,9].…”

Instance detection by keypoint matching beyond the nearest neighbor

“…The binary descriptors (such as [1,5,10,11,18,21,22]) are extensively used in real-time applications for object detection This work was supported by the TÜBİTAK project 113E496.…”