Large-scale visual sentiment ontology and detectors using adjective noun pairs

By Damian Borth, Rongrong Ji, Tao Chen, Thomas Breuel, Shih-Fu Chang. 2013

850 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Recognizing Image Style (2013, methodology): “[3] performed sentiment analysis on images using object classifiers trained on adjectivenoun pairs.”
Automatic Understanding of Image and Video Advertisements (2017, background): “This again shows that ads represent a new domain of images whose decoding requires novel techniques.”
DeepSentiBank: Visual Sentiment Concept Classification with Deep Convolutional Neural Networks (2014, methodology): “In [1] binary SVM classiﬁers of the ANPs are trained on the whole images, denoted as SentiBank 1.1.”
Sentiment Analysis: An Overview from Linguistics (2016, neutral): “Both Borth et al (2013) and Wang et al (2015) use a combination of characteristics of images posted online and text (comments and tags) about the image, to identify the sentiment conveyed by images.”
Real-time Analysis and Visualization of the YFCC100m Dataset (2015, background): “Over the last years different visual recognition tasks in computer vision and multimedia research have been introduced: object classification and detection [6, 5], semantic segmentation [10], concept detection [12], multimedia event detection [3], affective or emotional categorization [14], or visua”
Automatic Image Dataset Construction from Click-through Logs Using Deep Neural Network (2015, background): “As this step is quite time consuming and labor intensive, NUS-WIDE only partially labelled the whole dataset [3], while TinyImage [25] and visual sentiment dataset [1] keep all raw candidate images without manual labelling.”
What you Feel, Is what you like Influence of Message Appeals on Customer Engagement on Instagram (2020, methodology): “A number of approaches attempt to model visual valence and arousal (Borth et al. 2013; Chen et al. 2014). This study uses the Multilingual Visual Sentiment Ontology (MVSO) model since it represents the state of the art in terms of accuracy and scope (Jou et al. 2015). The MVSO model is trained on 7 ”
The YLI-MED Corpus: Characteristics, Procedures, and Plans (2015, background): “To manage the risk of user bias, creators of datasets from online platforms like Flickr and YouTube often set a maximum number of images or videos per user (e.g., Ulges et al. 2012; Borth et al. 2013); we elected to follow suit.”
Exploring Principles-of-Art Features For Image Emotion Recognition (2014, methodology): “Combining the visual concept detection method, such as SentiBank [4], may help to tackle this problem and further improve the emotion recognition performance.”
A survey of multimodal sentiment analysis (2017, methodology): “Whereas these were mostly noun concepts, in [14,107], Borth et al.”
Contextual Enrichment of Remote-Sensed Events with Social Media Streams (2016, background): “DeepSentiBank is a fine-tuned Convolutional Neural Network (CNN), that is based on the Visual Sentiment Ontology (VSO) [2].”
Query Expansion Techniques for Information Retrieval: a Survey (2017, methodology): “Reference [43] uses multitag for image retrieval, whereas [166] retrieves images using query adaptive hashing method.”
What makes an image popular? (2014, methodology): “Similarly, image quality and aesthetics are other attributes that have been recently explored in substantial detail [6, 8, 3].”
Multimodal Sentiment Analysis: A Survey of Methods, Trends, and Challenges (2023, background): “Borth et al. [88] developed a large-scale semantic concept visual sentiment ontology (VSO) named SentiBank [89] based on web mining and psychological theories.”
Action Modifiers: Learning From Adverbs in Instructional Videos (2019, background): “Learning adjectives for nouns has been investigated in the context of recognizing object-attribute pairs [4, 7, 20, 34, 36, 37, 50, 51] from images.”