Title :
Horror Image Recognition Based on Context-Aware Multi-Instance Learning
Author :
Bing Li ; Weihua Xiong ; Ou Wu ; Weiming Hu ; Maybank, Stephen ; Shuicheng Yan
Author_Institution :
Nat. Lab. of Pattern Recognition, Beijing, China
Abstract :
Horror content sharing on the Web is a growing phenomenon that can interfere with our daily life and affect the mental health of those involved. As an important form of expression, horror images have their own characteristics that can evoke extreme emotions. In this paper, we present a novel context-aware multi-instance learning (CMIL) algorithm for horror image recognition. The CMIL algorithm identifies horror images and picks out the regions that cause the sensation of horror in these horror images. It obtains contextual cues among adjacent regions in an image using a random walk on a contextual graph. Borrowing the strength of the fuzzy support vector machine (FSVM), we define a heuristic optimization procedure based on the FSVM to search for the optimal classifier for the CMIL. To improve the initialization of the CMIL, we propose a novel visual saliency model based on the tensor analysis. The average saliency value of each segmented region is set as its initial fuzzy membership in the CMIL. The advantage of the tensor-based visual saliency model is that it not only adaptively selects features, but also dynamically determines fusion weights for saliency value combination from different feature subspaces. The effectiveness of the proposed CMIL model is demonstrated by its use in horror image recognition on two large-scale image sets collected from the Internet.
Keywords :
content management; feature extraction; feature selection; fuzzy set theory; graph theory; image classification; learning (artificial intelligence); object detection; optimisation; random processes; support vector machines; tensors; CMIL algorithm; FSVM; Internet; context-aware multi-instance learning; contextual graph; feature selection; feature subspace; fusion weight determination; fuzzy membership; fuzzy support vector machine; heuristic optimization procedure; horror content sharing; horror image identification; horror image recognition; optimal classifier; random walk; region segmentation; tensor-based visual saliency model; Feature extraction; Image classification; Image recognition; Optimization; Support vector machines; Training; Visualization; Horror image recognition; context-aware multi-instance learning; horror image recognition; visual saliency;
Journal_Title :
Image Processing, IEEE Transactions on
DOI :
10.1109/TIP.2015.2479400