{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/deep-gaze-i-boosting-saliency-prediction-with","title":"Deep Gaze I: Boosting Saliency Prediction with Feature Maps Trained on ImageNet","arxiv_id":"1411.1045","date":"2014-11-04","proceeding":null,"authors":["Matthias Kümmerer","Lucas Theis","Matthias Bethge"],"abstract":"Recent results suggest that state-of-the-art saliency models perform far from\noptimal in predicting fixations. This lack in performance has been attributed\nto an inability to model the influence of high-level image features such as\nobjects. Recent seminal advances in applying deep neural networks to tasks like\nobject recognition suggests that they are able to capture this kind of\nstructure. However, the enormous amount of training data necessary to train\nthese networks makes them difficult to apply directly to saliency prediction.\nWe present a novel way of reusing existing neural networks that have been\npretrained on the task of object recognition in models of fixation prediction.\nUsing the well-known network of Krizhevsky et al. (2012), we come up with a new\nsaliency model that significantly outperforms all state-of-the-art models on\nthe MIT Saliency Benchmark. We show that the structure of this network allows\nnew insights in the psychophysics of fixation selection and potentially their\nneural implementation. To train our network, we build on recent work on the\nmodeling of saliency as point processes.","url_abs":"http://arxiv.org/abs/1411.1045v4","url_pdf":"http://arxiv.org/pdf/1411.1045v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"deep-gaze-i-boosting-saliency-prediction-with","repo_url":"https://github.com/matthias-k/DeepGaze","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"object-recognition","task_name":"Object Recognition"},{"task_slug":"point-processes","task_name":"Point Processes"},{"task_slug":"saliency-prediction","task_name":"Saliency Prediction"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1411.1045","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}