The modification of the descriptor in human detector using Histogram of Oriented Gradients and support vector machine is presented. The proposed modification requires inserting the average cell intensitiesresulting with the increase of the length of the descriptor from 3780 to 4200 values, but it is easy to compute and instantly gives 14-26% of miss rate improvement at 10^-4 False Positives Per Window (FPPW). The modification has been tested on two versions of HOG-based descriptors: classic Dalal-Triggs and the modified one, where, instead of spatial Gaussian masks for blocks, an additional central cell has been used. The proposed modificationis suitable for hardware implementations of HOG-based detectors, enabling to increase the detection accuracy or toresign from the use of some hardware-unfriendly operations, such as spatial Gaussian mask. The results of thetests of the influence on the intensity changes of test images are also presented.
Authors
Additional information
- Category
- Aktywność konferencyjna
- Type
- publikacja w wydawnictwie zbiorowym recenzowanym (także w materiałach konferencyjnych)
- Language
- angielski
- Publication year
- 2015
