Rationale and objectives: Computer-aided detection and diagnosis (CAD) systems have been developed in the past two decades to assist radiologists in the detection and diagnosis of lesions seen on breast imaging exams, thus providing a second opinion. Mammographic databases play an important role in the development of algorithms aiming at the detection and diagnosis of mammary lesions. However, available databases often do not take into consideration all the requirements needed for research and study purposes. This article aims to present and detail a new mammographic database.

Materials and methods: Images were acquired at a breast center located in a university hospital (Centro Hospitalar de S. João [CHSJ], Breast Centre, Porto) with the permission of the Portuguese National Committee of Data Protection and Hospital's Ethics Committee. MammoNovation Siemens full-field digital mammography, with a solid-state detector of amorphous selenium was used.

Results: The new database-INbreast-has a total of 115 cases (410 images) from which 90 cases are from women with both breasts affected (four images per case) and 25 cases are from mastectomy patients (two images per case). Several types of lesions (masses, calcifications, asymmetries, and distortions) were included. Accurate contours made by specialists are also provided in XML format.

Conclusion: The strengths of the actually presented database-INbreast-relies on the fact that it was built with full-field digital mammograms (in opposition to digitized mammograms), it presents a wide variability of cases, and is made publicly available together with precise annotations. We believe that this database can be a reference for future works centered or related to breast cancer imaging.

Rationale and objectives: Computer-aided detection and diagnosis (CAD) systems have been developed in the past two decades to assist radiologists in the detection and diagnosis of lesions seen on breast imaging exams, thus providing a second opinion. Mammographic databases play an important role in the development of algorithms aiming at the detection and diagnosis of mammary lesions. However, available databases often do not take into consideration all the requirements needed for research and study purposes. This article aims to present and detail a new mammographic database. Materials and methods: Images were acquired at a breast center located in a university hospital (Centro Hospitalar de S. João [CHSJ], Breast Centre, Porto) with the permission of the Portuguese National Committee of Data Protection and Hospital s Ethics Committee. MammoNovation Siemens full-field digital mammography, with a solid-state detector of amorphous selenium was used. Results: The new database-INbreast-h

INBREAST

## Overview

INBREAST is a dataset for object detection tasks - it contains Breast Mass TEka annotations for 387 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Inbreast

## Overview

Inbreast is a dataset for object detection tasks - it contains Breast 3OpO W7Rm annotations for 343 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [ODbL v1.0 license](https://creativecommons.org/licenses/ODbL v1.0).

The mammography dataset includes both benign and malignant tumors. In order to create the pictures for this dataset, 106 masses from the INbreast dataset, 53 masses from the MIAS dataset, and 2188 masses from the DDSM dataset were initially extracted. Then, we preprocess our photos using contrast-limited adaptive histogram equalization and data augmentation. Inbreast dataset has 7632 photos, MIAS dataset has 3816 images, and DDSM dataset includes 13128 images after data augmentation. Additionally, we combine DDSM, MIAS, and INbreast. The size of each image was changed to 227*227 pixels.

Dataset

This dataset was created by noura bentaher

Released under CC0: Public Domain

Contents

PurposeUsing computer-aided design (CAD) systems, this research endeavors to enhance breast cancer segmentation by addressing data insufficiency and data complexity during model training. As perceived by computer vision models, the inherent symmetry and complexity of mammography images make segmentation difficult. The objective is to optimize the precision and effectiveness of medical imaging.MethodsThe study introduces a hybrid strategy combining shape-guided segmentation (SGS) and M3D-neural cellular automata (M3D-NCA), resulting in improved computational efficiency and performance. The implementation of Shape-guided segmentation (SGS) during the initialization phase, coupled with the elimination of convolutional layers, enables the model to effectively reduce computation time. The research proposes a novel loss function that combines segmentation losses from both components for effective training.ResultsThe robust technique provided aims to improve the accuracy and consistency of breast tumor segmentation, leading to significant improvements in medical imaging and breast cancer detection and treatment.ConclusionThis study enhances breast cancer segmentation in medical imaging using CAD systems. Combining shape-guided segmentation (SGS) and M3D-neural cellular automata (M3D-NCA) is a hybrid approach that improves performance and computational efficiency by dealing with complex data and not having enough training data. The approach also reduces computing time and improves training efficiency. The study aims to improve breast cancer detection and treatment methods in medical imaging technology.

All code used in this study is available at: https://github.com/hwejin23/MAMMO_Retinanet. (ZIP)

Inbreast Class

## Overview

Inbreast Class is a dataset for object detection tasks - it contains Tumor annotations for 694 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Dataset

This dataset was created by cloner174

Released under CC0: Public Domain

Contents

INbreast BD

## Overview

INbreast BD is a dataset for classification tasks - it contains BD Level annotations for 410 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Dataset

This dataset was created by noura bentaher

Released under CC0: Public Domain

Contents

Performance comparison of the mass detection models.

The global breast imaging software market, currently valued at $479 million in 2025, is projected to experience robust growth, driven by a Compound Annual Growth Rate (CAGR) of 7.5% from 2025 to 2033. This expansion is fueled by several key factors. The increasing prevalence of breast cancer globally necessitates advanced diagnostic tools, leading to higher demand for sophisticated software solutions that improve accuracy, efficiency, and workflow in breast imaging centers and hospitals. Furthermore, technological advancements such as AI-powered image analysis, improved 3D and tomosynthesis capabilities, and the growing adoption of cloud-based platforms are contributing to market growth. The integration of these technologies allows for faster and more accurate detection of breast abnormalities, personalized treatment plans, and reduced healthcare costs in the long run. The market also benefits from rising awareness among women regarding regular breast cancer screening and increasing investments in healthcare infrastructure, especially in developing economies. Competition in the breast imaging software market is intensifying, with established players like Hologic, General Electric Company, and Carestream Health competing with emerging innovative companies such as DeepTek and Volpara Health. These companies are focusing on developing advanced algorithms, user-friendly interfaces, and comprehensive solutions that cater to the evolving needs of radiologists and healthcare professionals. While the market faces some restraints like high initial investment costs for advanced software and the need for extensive training for healthcare staff, the overall positive trajectory suggests significant growth potential across various segments and regions in the coming years. The market's segmentation includes different software types (e.g., CAD, PACS, AI-based analysis), deployment models (cloud-based, on-premise), and end-users (hospitals, imaging centers, clinics). North America and Europe are anticipated to maintain their dominant market share due to advanced healthcare infrastructure and high adoption rates of technologically advanced healthcare solutions. However, significant growth opportunities exist in Asia-Pacific and other developing regions as healthcare infrastructure expands and awareness around breast cancer screening increases.

Comparison between related works and the presented model based on the INbreast dataset.

Market Analysis for Artificial Intelligence in Breast Imaging The global market for artificial intelligence (AI) in breast imaging is expected to reach a value of XXX million by 2033, growing at a CAGR of XX% during the forecast period (2025-2033). Key drivers of this growth include the rising incidence of breast cancer, the increasing adoption of AI-powered imaging technologies, and the growing awareness of the benefits of AI in improving diagnostic accuracy and reducing patient waiting times. Major trends in the market include the development of cloud-based AI solutions and the integration of AI algorithms into existing breast imaging systems. Competitive Landscape and Outlook The competitive landscape of the AI in breast imaging market is characterized by the presence of both established players and emerging start-ups. Leading companies in the market include GE, CureMetrix, Densitas, QView Medical, IBM, Google, Icad, Philips, Amazon, Siemens, NVIDIA Corporation, Intel, Bayer (Blackford Analysis), Fujifilm, Aidoc, Arterys, Lunit, ContextVision AB, deepcOS, and Volpara Health Technologies Ltd. Key strategies adopted by these companies include product innovation, partnerships, and strategic acquisitions. The market is expected to witness continued growth in the coming years, driven by increasing investments in AI research and development and the adoption of AI-powered imaging solutions by healthcare providers worldwide.

This dataset contains data mapping questions addressed by the selected studies. (MQ7 has 3 sub-questions: D: dataset, V: validation method, M: metrics)

The data is presented in a pdf file: MappingQuestions.pdf. Contains the list of responses to the mapping questions.

Contact InformationFor further information or inquiries about this dataset, please contact [Juan Manuel Carrillo de Gea] at [jmcdg1@um.es].

In 2023, over 6.3 thousand cases of cancer were detected during the national breast cancer screening program. This was a decrease from the previous year, in which the number of screen-detected cancers was around seven thousand.

Source: Cancer Screening Database, HPA. Data by February 8, 2014. Note: 1. Follow-up rate for positive cases Number of positive cases completed follow-up / Number of positive cases. 2. Positive breast cancer cases are defined as: Breast screening results are "0, 4, 5", and screening dates are between October 1, 2012, and September 30, 2013.

Whole slide imaging of 396 full cases of axillary lymph nodes in breast cancer cases. Included are both sentinel node surgery and axillary dissections pre, peri or post breast cancer surgery or treatment. Sentinel node cases are cut in three levels (stained with HE) and one additional slide immunohistochemically stained with CKAE1/AE3. The number of sentinel node cases with complete immunohistochemical staining is 321. The axillary dissections are cut with one cut level as default. No frozen sections included. The cases are anonymised and exported from the digital archive at the Department of Clinical Pathology in Linköping, Region Östergötland. Included are both positive and negative cases. Some metadata on case level is available (positive or negative case, number of nodes, primary tumour and if neoadjuvant treatment in axillary dissections was given).