Automatic Gauze Tracking in Laparoscopic Surgery using Image Texture Analysis

Published: 5 Jan 2020 | Version 5 | DOI: 10.17632/yh2yxwbc9m.5
Contributor(s):

Description of this data

This dataset contain the code, images and video results of the analysed algorithms: LBP|VAR, CLBP and the convolutional neural network ResNet50.

Experiment data files

  • CLBP
    Cite

    Code CLBP

    • common
      Cite

      Code

    • include
      Cite
    • patternData
      Cite

      Contain the pattern histogram

    • processVideo
      Cite

      Main function for video processing. Contain the project for Code:Blocks

    • videos
      Cite
      • test
        Cite
  • CNN
    Cite

    Code CNN ResNet50

    • code
      Cite
    • img_dataset
      Cite
      • background
        Cite
      • gauze
        Cite
    • results
      Cite
      • blockValues
        Cite

        results/blockValues contains the test images with the percentage of probability overprinted that the block is gauze, according to the neuronal network.

  • LBPVAR
    Cite

    Code LBP|VAR

    • common
      Cite
    • include
      Cite
    • patternData
      Cite
    • processVideo
      Cite
    • videos
      Cite
      • test
        Cite
  • Results
    Cite

    This folder contains images, videos and tables with the results obtained using LBP, LBPVAR, CLBP and CNN approaches.

    • resultsCLBP
      Cite
      • images
        Cite

        Images with the results. In GREEN the blocks correctly classified as gauze (true positives), in RED the blocks wronly classified as gauze (false positives) and in YELLOW the blocks where the gauze has not been detected (false negatives).

        • CLBP
          Cite

          Results obtained on the suite of images with CLBP. No postprocessing.

        • finalCLBP
          Cite

          Results obtained on the suite of images with CLBP with postprocessing.

      • tables
        Cite
      • video
        Cite

        Detection results obtained with CLBP(with postprocessing) on three videos: with clean gauze, with stained gauze and with soaked gauze.

    • resultsCNN
      Cite
      • images
        Cite

        Images with the results. In GREEN the blocks correctly classified as gauze (true positives), in RED the blocks wronly classified as gauze (false positives) and in YELLOW the blocks where the gauze has not been detected (false negatives).

      • tables
        Cite
    • resultsLBP
      Cite
      • images
        Cite
      • tables
        Cite
      • video
        Cite

        Detection results obtained with the original LBP algorithm on three videos: with clean gauze, with stained gauze and with soaked gauze.

    • resultsLBPVar
      Cite
      • images
        Cite

        Images with the results. In GREEN the blocks correctly classified as gauze (true positives), in RED the blocks wronly classified as gauze (false positives) and in YELLOW the blocks where the gauze has not been detected (false negatives).

        • finalLBPVAR
          Cite
        • LBPVAR
          Cite
      • tables
        Cite
      • video
        Cite

        Detection results obtained with LBP|VAR (with postprocessing) on three videos: with clean gauze, with stained gauze and with soaked gauze.

  • testImages
    Cite

    Suite of images used to compare results among the three approaches

Steps to reproduce

LBP|VAR contains the code of the LBP |Variance algorithm for gauze detection. This code makes a final morphologycal postprocessing to improve the results.
CLBP contains the code for Completed LBP gauze detection. This code makes a final morphologycal postprocessing to improve the results.
To compile and run please mantain always the same relative position in the folders.
CNN contains the results and code for training and test the neural network. code/red_gasas.pt contains the neural network

Latest version

  • Version 5

    2020-01-05

    Published: 2020-01-05

    DOI: 10.17632/yh2yxwbc9m.5

    Cite this dataset

    de la Fuente, Eusebio (2020), “Automatic Gauze Tracking in Laparoscopic Surgery using Image Texture Analysis”, Mendeley Data, v5 http://dx.doi.org/10.17632/yh2yxwbc9m.5

Statistics

Views: 47550
Downloads: 985

Previous versions

Compare to version

Institutions

Universidad de Valladolid

Categories

Classification of Bioinformatical Subject

Licence

CC BY 4.0 Learn more

The files associated with this dataset are licensed under a Creative Commons Attribution 4.0 International licence.

What does this mean?
You can share, copy and modify this dataset so long as you give appropriate credit, provide a link to the CC BY license, and indicate if changes were made, but you may not do so in a way that suggests the rights holder has endorsed you or your use of the dataset. Note that further permission may be required for any content within the dataset that is identified as belonging to a third party.

Report