Skip to content

Latest commit

 

History

History
126 lines (89 loc) · 4.81 KB

Intel_MobileODT.md

File metadata and controls

126 lines (89 loc) · 4.81 KB

Intel & MobileODT Cervical Cancer Screening

Dataset Information

The Intel & MobileODT Cervical Cancer Screening dataset is an image-based medical imaging dataset aimed at identifying cervical types through cervical images. The dataset includes three types of cervical images: Type_1, Type_2, and Type_3.

  • Type 1: Exophytic cervix, fully visible, can be large or small.
  • Type 2: Has endocervical components, fully visible, may have exocervical components which can be large or small.
  • Type 3: Has endocervical components, not fully visible, may have exocervical components which can be large or small.

Labeled images are provided for the training set, while only images are provided for the testing set. This is a challenge initiated by Intel and MobileODT on the Kaggle platform. The clinical significance of this dataset lies in improving the early screening and treatment effectiveness of cervical cancer, especially in resource-limited areas. By developing algorithms that accurately identify cervical types, it enhances the quality of telemedicine, helps medical providers make more accurate treatment decisions, avoids ineffective treatments, and provides appropriate referrals for cases that require more advanced treatment. This not only promotes the development of medical AI but also expands medical coverage, helping women worldwide receive more effective cervical cancer prevention and treatment services.

Dataset Meta Information

Dimensions Modality Task Type Anatomical Structures Number of Categories Data Volume File Format
2D Hysteroscopy Classification Cervix 3 6733 JPG

Resolution Details

Dataset Statistics size
min [4128, 4160]
median [3096, 3264]
max [480, 480]

Label Information Statistics

Cervical Type Type 1 Type 2 Type 3
Number of Images 1191 3565 1976
Percentage 17.68% 52.94% 29.34%
Minimum Size [480, 480] [480, 640] [480, 640]
Median Size [3096, 4128] [3096, 4128] [2448, 3264]
Maximum Size [4128, 4160] [4128, 4160] [4128, 4160]

Visualization

Type 1.

Type 2.

Type 3.

File Structure

Dataset
├── train.7z
│   ├── Type_1
│   │   └── [training images for Type 1]
│   ├── Type_2
│   │   └── [training images for Type 2]
│   └── Type_3
│       └── [training images for Type 3]
├── test.7z
│   └── [all test images in one folder]
├── sample_submission.csv
├── fixed_labels_v2.csv
├── removed_files.csv
├── sample_submission_stg2.csv
├── solution_stg1_release.csv
├── test_stg2.7z
│   └── [stage 2 test images]
└── test_stg2_7z_password.txt

Authors and Institutions

BenO (Kaggle)

jljones (Kaggle)

Kumar H (Kaggle)

Meg Risdal (Kaggle)

MRao (Kaggle)

Vadim Sherman (Kaggle)

Vipul (Kaggle)

Wendy Kan (Kaggle)

Yau Ben-Or (Kaggle)

Source Information

Official Website: https://kaggle.com/competitions/intel-mobileodt-cervical-cancer-screening

Download Link: https://kaggle.com/competitions/intel-mobileodt-cervical-cancer-screening

Article Address: TBD

Publication Date: 2017-3

Citation

@misc{intel-mobileodt-cervical-cancer-screening,
    author = {BenO, jljones, Kumar H, Meg Risdal, MRao, Vadim Sherman, Vipul, Wendy Kan, Yau Ben-Or},
    title = {Intel & MobileODT Cervical Cancer Screening},
    publisher = {Kaggle},
    year = {2017},
    url = {https://kaggle.com/competitions/intel-mobileodt-cervical-cancer-screening}
}

Original introduction article is here.