Mastering Computer Vision with TensorFlow 2.x

The computational cost of LBP

LBP is computationally less expensive compared to the traditional neural network method. The computational cost of LBP was presented by Li Liu, Paul Fieguth, Xiaogang Wang, Matti Pietik äinen, and Dewen Hu in their paper Evaluation of LBP and Deep Texture Descriptors with A New Robustness Benchmark. The details of the paper can be found in the here: https://www.ee.cuhk.edu.hk/~xgwang/papers/liuFWPHeccv16.pdf

The authors determined the computational cost in terms of average time spent on the feature extraction of 480 images that were 128 x 128 in size on a 2.9 GHz Intel Quad-Core CPU with 16 GB RAM. The time does not include training time. The study found that LBP feature extraction was very fast compared to AlexNet ...