Work

L1-regularized Logistic Regression Stacking and Transductive CRF Smoothing for Action Recognition in Video