What makes an image memorable?

By Phillip Isola, Jianxiong Xiao, Antonio Torralba, Aude Oliva. 2011

473 citations recorded by the source corpus.

When glancing at a magazine, or browsing the Internet, we are continuously being exposed to photographs. Despite of this overflow of visual information, humans are extremely good at remembering thousands of pictures along with some of their visual details. But not all images are equal in memory. Some stitch to our minds, and other are forgotten. In this paper we focus on the problem of predicting how memorable an image will be. We show that memorability is a stable property of an image that is shared across different viewers. We introduce a database for which we have measured the probability that each picture will be remembered after a single view. We analyze image features and labels that contribute to making an image memorable, and we train a predictor based on global image descriptors. We find that predicting image memorability is a task that can be addressed with current computer vision techniques. Whereas making memorable images is a challenging task in visualization and photography, this work is a first attempt to quantify this useful quality of images.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

Recognizing Image Style (2013, background): “[12], who used various high-level features to predict human judgements of image memorability.”
Modifying the Memorability of Face Photographs (2013, methodology): “Metric Face database [2] Scene database [12] Human Pred Human Pred”
Protest Activity Detection and Perceived Violence Estimation from Social Media Images (2017, methodology): “This method has also been frequently used in the literature [20].”
What Makes a Photograph Memorable? (2014, unknown): “Index Terms—Scene understanding, image memorability, global image features, attributes ✦”
The Interestingness of Images (2013, methodology): “[14], extended in [13] to explore memorability.”
Large-scale visual sentiment ontology and detectors using adjective noun pairs (2013, background): “The closest that comes to sentiment analysis for visual content is the analysis of aesthetics [6, 15, 22], interestingness [12], and affect or emotions [13, 21, 37, 36].”
Deep Learning the City: Quantifying Urban Perception at a Global Scale (2016, background): “There is a growing body of literature on predicting the perceptual responses to images , such as aesthetics [21], memorability [22], interestingness [23], and virality [24].”
What Makes a Visualization Memorable? (2013, methodology): “These studies have demonstrated that the differences in the memorability of different images are consistent across observers, which implies that memorability is an intrinsic property of an image [21, 20].”
Summarization of Egocentric Videos: A Comprehensive Survey (2017, background): “Unlike for image memorability [54], for egocentric summa-”
DeepSentiBank: Visual Sentiment Concept Classification with Deep Convolutional Neural Networks (2014, background): “There is also work trying to model “non-conventional” concepts or properties of the images, such as image aesthetic and quality [16, 22], memorability [12], interestingness [12], and aﬀection/emotions [21, 35, 13, 21, 35, 37].”
Recognizing and Curating Photo Albums via Event-Specific Image Importance (2017, background): “Image importance is a complex image property which is related to various factors, such as aesthetics [15], interestingness [5] and image memorability [11].”
Salient object detection: A survey (2014, background): “r example, a detective sees a crime scene differently than a policeman or a pedestrian. Some related topics, closely or remotely, to visual saliency include: object importance [26]–[28], memorability [29], scene clutter [30], video interestingness [31]–[34], surprise [35], image quality assessment [”
What makes an image popular? (2014, background): “One such attribute is image memorability which has been shown to be an intrinsic image property [21] with diﬀerent image regions contributing diﬀerently to an image’s memorability [29, 28].”
The intrinsic memorability of face photographs. (2013, methodology): “We conducted a large-scale visual memory experiment using online participants (877 workers) from the crowdsourcing tool Amazon Mechanical Turk (AMT), following the protocol of Isola et al. (2011a).”
Predicting Failures of Vision Systems (2014, background): “The subsequent " system " may even be a user: an image memorability predictor [25] may be used by a graphics designer.”
Adopting Abstract Images for Semantic Scene Understanding (2016, background): “This is in contrast with tests for visual change detection [76] that show the target and altered image consecutively one after the other.”
Streetscore -- Predicting the Perceived Safety of One Million Streetscapes (2014, background): “Research in visual analysis beyond semantics, such as interestingness [5], beauty [4] and memorability [10], has been a recent topic of interest in computer vision.”
The Synthesizability of Texture Examples (2014, background): “Even if challenging, the task may be doable, given that other qualitative image characteristics could be quantified, like quality [26], memorability [14], or interestingness [11].”
Interestingness Prediction by Robust Learning to Rank (2014, background): “Early efforts on image interestingness prediction focus on different aspects than interestingness as such, including image quality [22], memorability [19], and aesthetics [7].”