{"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/multimodal-task-driven-dictionary-learning","title":"Multimodal Task-Driven Dictionary Learning for Image Classification","arxiv_id":"1502.01094","date":"2015-02-04","proceeding":null,"authors":["Soheil Bahrampour","Nasser M. Nasrabadi","Asok Ray","W. Kenneth Jenkins"],"abstract":"Dictionary learning algorithms have been successfully used for both\nreconstructive and discriminative tasks, where an input signal is represented\nwith a sparse linear combination of dictionary atoms. While these methods are\nmostly developed for single-modality scenarios, recent studies have\ndemonstrated the advantages of feature-level fusion based on the joint sparse\nrepresentation of the multimodal inputs. In this paper, we propose a multimodal\ntask-driven dictionary learning algorithm under the joint sparsity constraint\n(prior) to enforce collaborations among multiple homogeneous/heterogeneous\nsources of information. In this task-driven formulation, the multimodal\ndictionaries are learned simultaneously with their corresponding classifiers.\nThe resulting multimodal dictionaries can generate discriminative latent\nfeatures (sparse codes) from the data that are optimized for a given task such\nas binary or multiclass classification. Moreover, we present an extension of\nthe proposed formulation using a mixed joint and independent sparsity prior\nwhich facilitates more flexible fusion of the modalities at feature level. The\nefficacy of the proposed algorithms for multimodal classification is\nillustrated on four different applications -- multimodal face recognition,\nmulti-view face recognition, multi-view action recognition, and multimodal\nbiometric recognition. It is also shown that, compared to the counterpart\nreconstructive-based dictionary learning algorithms, the task-driven\nformulations are more computationally efficient in the sense that they can be\nequipped with more compact dictionaries and still achieve superior performance.","url_abs":"http://arxiv.org/abs/1502.01094v2","url_pdf":"http://arxiv.org/pdf/1502.01094v2.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":"multimodal-task-driven-dictionary-learning","repo_url":"https://github.com/soheilb/multimodal_dictionary_learning","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"dictionary-learning","task_name":"Dictionary Learning"},{"task_slug":"face-recognition","task_name":"Face Recognition"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1502.01094","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}