AMNet: Memorability Estimation with Attention
This work addresses the challenge of predicting which images are more memorable for applications in advertising and media, but it is incremental as it builds on existing deep learning methods with an attention mechanism.
The paper tackled the problem of estimating image memorability using a deep neural network with visual attention, achieving state-of-the-art performance on SUN Memorability and LaMem datasets by closely matching human consistency in Spearman's rank correlation and mean squared error.
In this paper we present the design and evaluation of an end-to-end trainable, deep neural network with a visual attention mechanism for memorability estimation in still images. We analyze the suitability of transfer learning of deep models from image classification to the memorability task. Further on we study the impact of the attention mechanism on the memorability estimation and evaluate our network on the SUN Memorability and the LaMem datasets. Our network outperforms the existing state of the art models on both datasets in terms of the Spearman's rank correlation as well as the mean squared error, closely matching human consistency.