Photo from academic.microsoft.com
In this Letter, the authors put forward a loose classifier to determine whether the content of a given image belongs to screen content or natural scene. In the proposed classifier,… Click to show full abstract
In this Letter, the authors put forward a loose classifier to determine whether the content of a given image belongs to screen content or natural scene. In the proposed classifier, they first extract 14 features by means of three typical histograms, the greyscale image histogram, local binary pattern histograms, and histograms of the oriented gradient, respectively. Then, in the second step, they try to build a non-linear mapping, with stacking-based integrated learners, from extracted features to the classification label. Experimental results prove that the effectiveness of their proposed method as compared with popular algorithms.
Journal Title: Electronics Letters
Year Published: 2019
Link to full text (if available)
Share on Social Media: Sign Up to like & get
recommendations!