[data request] OASBUD

[sputney13]

• Name of dataset: OASBUD (Open Access Series of Breast Ultrasound Data)
• URL of dataset: https://zenodo.org/record/545928#.X0xKf8hKg2z
• License of dataset: Creative Commons 4.0
• Short description of dataset and use case(s): This small dataset contains 200 ultrasound scans (2 orthogonal scans each) of 52 malignant and 48 benign breast tumors. The scans are stored as rf data, and each scan includes a same-size mask that denotes the region-of-interest for the tumor. The 100 tumors were ranked on the BI-RADS scale, which describes the probability of lesion malignancy, and classified as malignant or benign based on histological assessment of a core-needle biopsy or (for some benign cases) two years of observation by a radiologist with 18 years of experience in the ultrasonic examination of breast lesions. There are two key use cases for the dataset: first, using the rf data to estimate quantitative ultrasound parameters that may be the best for tumor classification, and second, in developing computer-aided diagnosis systems for breast tumors (involving image processing, segmentation, and classification tasks).

Folks who would also like to see this dataset in tensorflow/datasets, please thumbs-up so the developers can know which requests to prioritize.

And if you'd like to contribute the dataset (thank you!), see our guide to adding a dataset.

I will be working on adding this dataset myself.

[Liqq1]

Hi, what should I do with the .mat file? Why are the elements in the matrix I read negative?😭

[sputney13]

Hi, what should I do with the .mat file? Why are the elements in the matrix I read negative?😭

Hello, the data in the .mat matrix is from a raw rf ultrasound scan. It is not unexpected that the values would be negative. Even after post-processing using the methods added in the pull request (and recommended by the authors of the dataset), values will still be negative. You can still visualize these negative values using matplotlib.

For an example of how to download, process, and visualize the data, you can reference this Google Colab notebook: https://colab.research.google.com/drive/10qm-Jy3AX7tY6ZMCrJEW4QbWlvPgXsQ_?usp=sharing

[Liqq1]

Thank you for your reply Where can I find 'oasbud/raw_rf' folders and 'oasbud/b_mode'folders ?

------------------ 原始邮件 ------------------ 发件人: "tensorflow/datasets" @.>; 发送时间: 2022年7月14日(星期四) 上午6:58 @.>; @.@.>; 主题: Re: [tensorflow/datasets] [data request] OASBUD (#2371)

Hi, what should I do with the .mat file? Why are the elements in the matrix I read negative?😭

Hello, the data in the .mat matrix is from a raw rf ultrasound scan. It is not unexpected that the values would be negative. Even after post-processing using the methods added in the pull request (and recommended by the authors of the dataset), values will still be negative. You can still visualize these negative values using matplotlib.

For an example of how to download, process, and visualize the data, you can reference this Google Colab notebook: https://colab.research.google.com/drive/10qm-Jy3AX7tY6ZMCrJEW4QbWlvPgXsQ_?usp=sharing

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

[sputney13]

Thank you for your reply Where can I find 'oasbud/raw_rf' folders and 'oasbud/b_mode'folders ?

I'm not sure where on your machine TFDS would download the data to (as those are just intermediary folders for the raw and post-processed data), but if you're looking to access that data directly you might as well just download the .mat file from the original research. The point of accessing the data through TFDS is to abstract the data and the tools for manipulating it from whatever training you are doing. If you need an example of how to pass the data loaded in from TFDS into a model for training, you can consult this notebook: https://colab.research.google.com/drive/1Lak9TGaZ_Nvk3ozqflLs6pkC5LmG5fcR?usp=sharing