ABSTRACT: –

The main reason for death in women is breast cancer. If most cancers are detected at an early stage, then the chances of survival are very high. In the body, new cells take over the area of old cells with the aid of an orderly boom as vintage cells die out. The machine of mutation controls the activation of genes in cells. As a result, the cells gain the ability to divide indefinitely, resulting in the production of extra cells similar to them and the formation of a tumor. This tumour may be benign or malignant. Benign tumours aren’t dangerous to one’s health, whereas malignant tumours are. The unchecked malignant tumours have the potential to unfold in different elements of the frame. Breast cancer detection is a complicated procedure. As a result, computer-aided analysis of breast cancers aids physician decision-making. The system for breast cancer detection has advanced with the use of backpropagation neural networks, and we examine its outcomes with the radial foundation function community. After accomplishing research, we determined that backpropagation neural communities are an awesome way to find out about breast cancers. Early detection of sickness has emerged as an important problem in recent years due to the fast population boom visible in clinical studies. The chance of death from breast cancer increases dramatically as the sector’s population continues to increase at an alarming rate. As compared to other cancers discovered to this point, breast cancer is the second most severe. Further to assisting a scientific group of workers in disorder prognosis, an automatic disorder detection system additionally gives dependable, effective, and speedy intervention, which reduces the probability of mortality. For this reason, there is a desire for efficient strategies that diagnose cancerous cells without human involvement and with high accuracy. In this research, photo processing techniques were used to broaden imaging biomarkers through mammography evaluation based totally on synthetic intelligence technology, aiming to catch breast cancer in its early stages to assist the prognosis and prioritisation of high-danger patients. A generalised regression synthetic neural network was developed and tested for its ability to differentiate between malignant and benign tumours on mammograms, yielding an accuracy of Breast cancer is one of the most common cancers that kills a large number of women. The excessive prevalence and mortality of breast cancer are due to its drastically low accuracy of prognosis. In this paper, we explore gadget learning models that can be implemented to help increase the accuracy of the prognosis of breast cancer. Breast cancer has emerged as the nemesis that threatens girls’ fitness worldwide. A number of research projects were challenged by this illness to improve the accuracy of its analysis and detection. However, the situation remains one of the most common lifestyle-threatening illnesses, threatening the lives of one in every six women. The anomaly of the purpose of this ailment will increase the problem of coping with such a disorder since it makes the preventive technique impossible. Therefore, the only remaining hope that the targets of this ailment have is early detection. Using picture processing techniques, this paper will assist in providing an accurate technique to come across the ailment at its toddler level. This file contains an in-depth explanation of the steps associated with image processing: photo enhancement, segmentation, and function extraction, which are routinely performed alongside the type of abnormality using a convolutional neural network (CNN).

SYSTEM:-

breast cancer is a leading cause of death among women worldwide and early detection is critical for successful treatment a neural network system can be an effective tool for breast cancer detection as it can analyze large amounts of data and identify patterns that may not be visible to the human eye

heres a high-level overview of a potential system for breast cancer detection using neural networks

data collection: the first step in developing a breast cancer detection system is to collect data this data could include mammography images patient demographics and other relevant medical information the data must be labeled as either positive indicating the presence of cancer or negative indicating no cancer

data preprocessing: once the data is collected it must be preprocessed to prepare it for analysis this could involve image resizing normalization and feature extraction feature extraction could involve identifying specific characteristics of the images that are indicative of breast cancer such as the presence of masses or calcifications

neural network architecture selection: there are many different neural network architectures that could be used for breast cancer detection including convolutional neural networks cnns and deep belief networks the choice of architecture will depend on the specific requirements of the project

Training the neural network :once the architecture is selected the neural network must be trained using the preprocessed data during the training phase the network will learn to recognize patterns in the data that are indicative of breast cancer.

validation and testing : after the neural network is trained it must be validated and tested to ensure that it is accurate and reliable this could involve testing the network on a separate dataset to ensure that it can correctly classify positive and negative cases.

integration and deployment: once the neural network is validated and tested it can be integrated into a larger system for breast cancer detection this system could be used by healthcare professionals to assist in the overall diagnosis of breast cancer .

a breast cancer detection system using neural networks can be a powerful tool for the early detection of breast cancer but it requires careful data collection preprocessing architecture selection training validation and testing to ensure that the system is accurate and reliable

PROPOSED SYSTEM:-

The marked framework for VGG16 we used in the proposed system It has 41 disturbed layers, which include 16 weight layers, 13 convolutional layers (Conv.), and 3 FC layers. VGG16 employs a small 3×3 kernel (filter) on all conversion layers with stride one. Max pooling layers always follow conversion layers. The input for VGG16 is fixed at 224×224 three-channel images. In VGG16, the three FC layers have different depths. The first two have the same channel size of 4096, while the last FC has a channel size of 1000, representing the number of the class marker in the imageNet dataset. The output layer is the soft maximum layer, which is responsible for the given probability for the input image. As with any pre-trained model, VGG16 requires heavy training if the weights are initialised randomly. So, in general, CCN models use transfer learning (TL) techniques. Transfer learning is a medium in which a model trained on one task is applied to an alternate analogous task in some way. I.e., we train a CNN model on an analogous problem to the problem that’s being addressed, where the input is the same but the output may be different in nature. In this case, the VGG 16 model is trained using the ImageNet dataset, which contains numerous real-world object images. Layer weights are also moved to a breast cancer classification task or feature extraction. Therefore, the training time is reduced. Either way, TL is a more important classification approach when a small dataset is estimated. Following the conformation of some layers from the pre-trained model, TL can be used for classification or feature extraction. In this study, the capability of the VGG16 with transfer literacy is applied to extract high-level features from the input images.

MODULES:-

• data collection module: this module will collect mammography images patient demographics and other relevant medical information the data must be labeled as either positive or negative
• data preprocessing module: this module will preprocess the collected data to prepare it for analysis this could involve image resizing normalization and feature extraction feature extraction could involve identifying specific characteristics of the images that are indicative of breast cancer such as the presence of masses or calcifications
• neural network architecture selection module: this module will choose the appropriate neural network architecture for breast cancer detection it will select from different neural network architectures such as cnns and dbns based on the specific requirements of the project
• training module :this module will train the neural network using the preprocessed data during the training phase the network will learn to recognize patterns in the data that are indicative of breast cancer .
• validation and testing module: this module will validate and test the trained neural network to ensure that it is accurate and reliable this could involve testing the network on a separate dataset to ensure that it can correctly classify positive and negative cases.
• user interface module: this module will provide a user interface that allows healthcare professionals to upload mammography images and view the results of the breast cancer detection system .
• reporting module: this module will generate reports on the results of the breast cancer detection system including the accuracy of the system and any potential false positives or false negatives
• integration and deployment module: this module will integrate the different modules into a complete breast cancer detection system and deploy it to a production environment where healthcare professionals can use it to assist in the diagnosis of breast cancer

APPLICATION:-

• Early detection of breast cancer: One of the primary applications of a breast cancer detection system using neural networks is early detection of breast cancer. Early detection is critical for successful treatment, and a neural network system can help healthcare professionals detect breast cancer at an earlier stage than they might be able to otherwise.

• Screening mammography interpretation: A breast cancer detection system using neural networks can help healthcare professionals interpret screening mammography results more accurately and efficiently. This can reduce the number of false positives and false negatives, leading to more accurate diagnoses.

• Personalized medicine: A breast cancer detection system using neural networks can be used to develop personalized treatment plans for patients with breast cancer. By analyzing data from a patient’s mammography images and medical history, the system can help healthcare professionals develop treatment plans tailored to the patient’s specific needs.

• Radiology education: A breast cancer detection system using neural networks can be used to educate radiologists and other healthcare professionals on the characteristics of breast cancer and how to interpret mammography results accurately.

• Research: A breast cancer detection system using neural networks can be used in research studies to analyze large amounts of data and identify patterns that may not be visible to the human eye. This can lead to new insights into the causes and treatment of breast cancer.

• Overall, a breast cancer detection system using neural networks has numerous applications in healthcare, from early detection of breast cancer to personalized medicine to radiology education and research. By leveraging the power of neural networks, healthcare professionals can improve the accuracy and efficiency of breast cancer diagnosis and treatment.

HARDWARE AND SOFTWARE REQUIREMENTS:-

HARDWARE:-

• Processor: i3 ,i5 or more
• RAM: 4GB or more
• Hard disk: 16 GB or more

SOFTWARE:-

• Operating System : Windows 10, 7, 8.
• Python
• anaconda
• Spyder,
• Jupyter notebook, Flask.
• MYSQL