Embracing New Techniques in Deep Learning for Estimating Image Memorability
This work addresses the need for more accurate and generalizable image memorability predictions for memory researchers, but it is incremental as it applies existing residual network techniques to this specific domain.
The authors tackled the problem of predicting image memorability by proposing five new deep learning models based on residual neural networks, which outperformed the prior state-of-the-art model that was found to be overfit, leading to improved generalizability across image categories.
Various work has suggested that the memorability of an image is consistent across people, and thus can be treated as an intrinsic property of an image. Using computer vision models, we can make specific predictions about what people will remember or forget. While older work has used now-outdated deep learning architectures to predict image memorability, innovations in the field have given us new techniques to apply to this problem. Here, we propose and evaluate five alternative deep learning models which exploit developments in the field from the last five years, largely the introduction of residual neural networks, which are intended to allow the model to use semantic information in the memorability estimation process. These new models were tested against the prior state of the art with a combined dataset built to optimize both within-category and across-category predictions. Our findings suggest that the key prior memorability network had overstated its generalizability and was overfit on its training set. Our new models outperform this prior model, leading us to conclude that Residual Networks outperform simpler convolutional neural networks in memorability regression. We make our new state-of-the-art model readily available to the research community, allowing memory researchers to make predictions about memorability on a wider range of images.