Visual Sentiment Prediction Based on Automatic Discovery of Affective Regions

TLDR

Automatic sentiment assessment from visual content has attracted attention as people increasingly express opinions via images and videos, yet labeling affective regions is subjective and labor‑intensive, making image‑level approaches desirable. The study investigates visual sentiment analysis and proposes to exploit local regions, motivated by the observation that both whole images and local regions convey significant sentiment information. The authors propose a framework that generates objectness‑based region candidates, selects reliable ones, feeds each through a CNN to compute sentiment scores, automatically discovers affective regions, and aggregates local and global CNN outputs for final predictions. The framework, requiring only image‑level labels, reduces annotation effort and outperforms state‑of‑the‑art methods on eight benchmark datasets.

Abstract

Automatic assessment of sentiment from visual content has gained considerable attention with the increasing tendency of expressing opinions via images and videos online. This paper investigates the problem of visual sentiment analysis, which involves a high-level abstraction in the recognition process. While most of the current methods focus on improving holistic representations, we aim to utilize the local information, which is inspired by the observation that both the whole image and local regions convey significant sentiment information. We propose a framework to leverage affective regions, where we first use an off-the-shelf objectness tool to generate the candidates, and employ a candidate selection method to remove redundant and noisy proposals. Then, a convolutional neural network (CNN) is connected with each candidate to compute the sentiment scores, and the affective regions are automatically discovered, taking the objectness score as well as the sentiment score into consideration. Finally, the CNN outputs from local regions are aggregated with the whole images to produce the final predictions. Our framework only requires image-level labels, thereby significantly reducing the annotation burden otherwise required for training. This is especially important for sentiment analysis since sentiment can be abstract, and labeling affective regions is too subjective and labor-consuming. Extensive experiments show that the proposed algorithm outperforms the state-of-the-art approaches on eight popular benchmark datasets.