PROJECT TITLE:

Semi-Supervised Image Classification Based on Local and Global Regression

ABSTRACT:

The insufficiency of labeled samples may be a major drawback in automatic image annotation. However, unlabeled samples are readily available and abundant. Hence, Semi-Supervised Learning ways, which utilize partly labeled samples and a large amount of unlabeled samples, have attracted increased attention in the field of image classification. Throughout the past decade, graph-based semi-supervised learning became one in all the foremost vital research areas in semi-supervised learning. In this letter, we tend to propose a novel and effective graph primarily based semi-supervised learning technique for image classification. The new method is predicated on native and international regression regularization. The native regression regularization adopts a group of native classification functions to preserve each native discriminative and geometrical data; whereas the global regression regularization preserves the global discriminative info and calculates the projection matrix for out-of-sample extrapolation. Intensive simulations primarily based on synthetic and real-world datasets verify the effectiveness of the proposed technique.