Conference paper Open Access

Deep Multi-task Learning with Label Correlation Constraint for Video Concept Detection

Markatopoulou, Foteini; Mezaris, Vasileios; Patras, Ioannis


JSON Export

{
  "files": [
    {
      "links": {
        "self": "https://zenodo.org/api/files/4dac3875-e1f0-420f-b40e-f5319059d173/mm16_1_preprint.pdf"
      }, 
      "checksum": "md5:a715e3e8d926a165019b1b350465caa1", 
      "bucket": "4dac3875-e1f0-420f-b40e-f5319059d173", 
      "key": "mm16_1_preprint.pdf", 
      "type": "pdf", 
      "size": 433735
    }
  ], 
  "owners": [
    22750
  ], 
  "doi": "10.1145/2964284.2967271", 
  "stats": {
    "version_unique_downloads": 64.0, 
    "unique_views": 126.0, 
    "views": 133.0, 
    "version_views": 133.0, 
    "unique_downloads": 64.0, 
    "version_unique_views": 126.0, 
    "volume": 28192775.0, 
    "version_downloads": 65.0, 
    "downloads": 65.0, 
    "version_volume": 28192775.0
  }, 
  "links": {
    "doi": "https://doi.org/10.1145/2964284.2967271", 
    "latest_html": "https://zenodo.org/record/162404", 
    "bucket": "https://zenodo.org/api/files/4dac3875-e1f0-420f-b40e-f5319059d173", 
    "badge": "https://zenodo.org/badge/doi/10.1145/2964284.2967271.svg", 
    "html": "https://zenodo.org/record/162404", 
    "latest": "https://zenodo.org/api/records/162404"
  }, 
  "created": "2016-10-21T16:52:02.102447+00:00", 
  "updated": "2020-01-20T16:59:25.014680+00:00", 
  "conceptrecid": "657412", 
  "revision": 7, 
  "id": 162404, 
  "metadata": {
    "access_right_category": "success", 
    "doi": "10.1145/2964284.2967271", 
    "description": "<p>In this work we propose a method that integrates multi-task learning (MTL) and deep learning. Our method appends a MTL-like loss to a deep convolutional neural network, in order to learn the relations between tasks together at the same time, and also incorporates the label correlations between pairs of tasks. We apply the proposed method on a transfer learning scenario, where our objective is to fine-tune the parameters of a network that has been originally trained on a large-scale image dataset for concept detection, so that it be applied on a target video dataset and a corresponding new set of target concepts. We evaluate the proposed method for the video concept detection problem on the TRECVID 2013 Semantic Indexing dataset. Our results show that the proposed algorithm leads to better concept-based video annotation than existing state-of-the-art methods.</p>", 
    "license": {
      "id": "CC-BY-4.0"
    }, 
    "title": "Deep Multi-task Learning with Label Correlation Constraint for Video Concept Detection", 
    "relations": {
      "version": [
        {
          "count": 1, 
          "index": 0, 
          "parent": {
            "pid_type": "recid", 
            "pid_value": "657412"
          }, 
          "is_last": true, 
          "last_child": {
            "pid_type": "recid", 
            "pid_value": "162404"
          }
        }
      ]
    }, 
    "communities": [
      {
        "id": "ecfunded"
      }, 
      {
        "id": "invid-h2020"
      }, 
      {
        "id": "moving-h2020"
      }
    ], 
    "grants": [
      {
        "code": "693092", 
        "links": {
          "self": "https://zenodo.org/api/grants/10.13039/501100000780::693092"
        }, 
        "title": "Training towards a society of data-savvy information professionals to enable open leadership innovation", 
        "acronym": "MOVING", 
        "program": "H2020", 
        "funder": {
          "doi": "10.13039/501100000780", 
          "acronyms": [], 
          "name": "European Commission", 
          "links": {
            "self": "https://zenodo.org/api/funders/10.13039/501100000780"
          }
        }
      }, 
      {
        "code": "687786", 
        "links": {
          "self": "https://zenodo.org/api/grants/10.13039/501100000780::687786"
        }, 
        "title": "In Video Veritas \u2013 Verification of Social Media Video Content for the News Industry", 
        "acronym": "InVID", 
        "program": "H2020", 
        "funder": {
          "doi": "10.13039/501100000780", 
          "acronyms": [], 
          "name": "European Commission", 
          "links": {
            "self": "https://zenodo.org/api/funders/10.13039/501100000780"
          }
        }
      }
    ], 
    "keywords": [
      "Concept detection; deep learning; video analysis"
    ], 
    "publication_date": "2016-10-17", 
    "creators": [
      {
        "affiliation": "CERTH", 
        "name": "Markatopoulou, Foteini"
      }, 
      {
        "affiliation": "CERTH", 
        "name": "Mezaris, Vasileios"
      }, 
      {
        "affiliation": "QMUL", 
        "name": "Patras, Ioannis"
      }
    ], 
    "meeting": {
      "dates": "October 2016", 
      "place": "Amsterdam", 
      "title": "ACM Multimedia"
    }, 
    "access_right": "open", 
    "resource_type": {
      "subtype": "conferencepaper", 
      "type": "publication", 
      "title": "Conference paper"
    }
  }
}
133
65
views
downloads
Views 133
Downloads 65
Data volume 28.2 MB
Unique views 126
Unique downloads 64

Share

Cite as