We explore how to leverage the performance of face feature points detection on mobile terminals from3 aspects. First, we optimize\nthe models used in SDM algorithms via PCA and Spectrum Clustering. Second, we propose an evaluation criterion using Linear\nDiscriminative Analysis to choose the best local feature descriptions which plays a critical role in feature points detection. Third, we\ntake advantage ofmulticore architecture of mobile terminal and parallelize the optimized SDMalgorithm to improve the efficiency\nfurther.The experiment observations show that our final accomplished GPC-SDM (improved Supervised Descent Method using\nspectrum clustering, PCA, andGPU acceleration) suppresses the memory usage, which is beneficial and efficient to meet the realtime\nrequirements.
Loading....