A novel training method has been proposed for increasing efficiency and generalization of support vector machine (SVM). The efficiency of SVM in classification is directly determi...
Visual codebook based quantization of robust appearance descriptors extracted from local image patches is an effective means of capturing image statistics for texture analysis and...
Abstract: This paper presents a novel approach for object classification and pose estimation which employs spherical light field rendering to generate virtual views based on synthe...
Previous Multiple Kernel Learning approaches (MKL) employ different kernels by their linear combination. Though some improvements have been achieved over methods using single kerne...
One of the obstacles in automatic program proving is to obtain suitable loop invariants. The invariant of a loop is a weakened form of its postcondition (the loop's goal, als...