Sample DATA4200 Data Acquisition and Management Report 2

DATA4200 Data Acquisition and Management Report 2 Sample

Your Task

This report will enable you to practice your LO1 and LO2 skills.

• LO1: Evaluate ethical data acquisition and best practice about project initiation

• LO2: Evaluate options for storing, accessing, distributing, and updating data during the life of a project.

• Complete all parts below. Consider the rubric at the end of the assignment for guidance on structure and content.

• Submit the results as a Word file in Turnitin by the due date.

Background

Unstructured data has typically been difficult to manage, since it has no predefined data model, is not always organised, may comprise multiple types. For example, data from thermostats, sensors, home electronic devices, cars, images, sounds and pdf files.

Given these characteristics, special collection, storage, and analysis methods, as well as software, have been created to take advantage of unstructured data.

Assessment Instructions

Given the considerations above, select one of the following industries for your assessment.

• Healthcare

• Retail - clothing

• Social Media

• Education

• Motor vehicles

• Fast Foods

1. Read relevant articles on the industry you have chosen.

2. Choose one application from that industry.

3. Introduce the industry and application, e.g., healthcare and image reconstruction.

4. Explain what sort of unstructured data could be used by an AI or Machine Learning algorithm in the area you chose.

a. Discuss best practice and options for

b. Accessing/collecting

c. Storing

d. Sharing

e. Documenting

f. and maintenance of the data

5. Propose a question that could be asked in relation to your unstructured data and what software might help you to run AI and answer the question.

Solution

Introduce the industry and application

The healthcare industry is made up of a variety of medical services, technologies, and professionals who work to improve people's health and well-being. It includes hospitals, clinics, pharmaceutical companies, manufacturers of medical devices, research institutions, and many more. For assignment help, The healthcare industry is always looking for new ways to improve patient care and outcomes in order to diagnose, treat, and prevent diseases.

One significant application inside the medical services industry is processing of medical images. The process of acquiring, analyzing, and interpreting medical images for the purposes of diagnosis and treatment is known as medical image processing. It is essential in a number of medical fields, including orthopedics, cardiology, neurology, oncology, and radiology. Medical images can be obtained through modalities like X-ray, computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and positron emission tomography (PET), and advanced algorithms and computer-based techniques are used in medical image processing to extract meaningful information from medical images. The human body's internal structures, organs, tissues, and physiological processes are depicted in great detail in these images. Patients and healthcare professionals alike can reap numerous advantages from using medical image processing. It empowers more precise and proficient determination by giving nitty gritty bits of knowledge into the presence, area, and attributes of anomalies or infections. Images can be analyzed by doctors to find conditions like tumors, fractures, and blocked blood vessels, which can help with treatment planning and monitoring (Diène et al., 2020). Medical image processing aids in the development of healthcare research and development. It makes it possible to create massive image databases for the purpose of training machine learning algorithms. This can help automate tasks related to image analysis, increase productivity, and cut down on human error. Besides, it supports the investigation of new imaging procedures, for example, useful X-ray or dissemination tensor imaging, which gives important bits of knowledge into cerebrum capabilities and brain network.

Discussion on sort of unstructured data could be used by an AI or Machine Learning algorithm in the processing of medical image

There are many different kinds of unstructured data that can be used in image processing. Information that does not adhere to a predetermined data model or organization is referred to as unstructured data. Here are a few instances of unstructured data utilized in the processing of medical image:

Image Pixels: Unstructured data is created from an image's raw pixel values. Algorithms can use the color information in each pixel, such as RGB values, to extract features or carry out tasks like image classification and object detection.

Metadata for Images: Metadata that accompanies images typically contains additional information about the image. The camera's make and model, exposure settings, GPS coordinates, timestamps, and other information may be included in this metadata (Galetsi et al., 2020). This information can be used by machine learning algorithms to improve image analysis, such as locating an image or adjusting for particular camera characteristics.

Figure 1: Machine learning for medical image processing
(Source: https://pubs.rsna.org)

Captions or descriptions of images: Human-created portrayals or subtitles related to pictures give text based settings that can be utilized in artificial intelligence calculations. For tasks like image search, content recommendation, or sentiment analysis, natural language processing techniques can analyze these descriptions and extract useful information.

Labels and annotations: Unstructured information can likewise incorporate manual comments or marks that are added to pictures by people. These annotations may indicate the presence of bounding boxes, semantic segmentation, regions of interest, or objects. AI calculations can involve this marked information for preparing and approval purposes, empowering assignments like article acknowledgment, semantic division, or picture restriction.

Image Content: Textual elements, such as signs, labels, or captions, can also be present in unstructured data contained within images (Panesar, 2019). Algorithms can process and analyze the textual information in these images thanks to the ability of optical character recognition (OCR) techniques to extract the text from the images.

Picture Setting: Unstructured data can be used to access information about an image's context, such as its source website, related images, or user interactions. Machine learning algorithms can improve content filtering, image comprehension and recommendation systems by taking the context into account.

Discuss Best Practice and Options

Accessing/collecting, Storing, Sharing, Documenting and maintenance of the data are very important for the healthcare industry. Here is the discussion on some options and practices related to these procedures in the healthcare industry and image processing.

Accessing/collecting

Collection of healthcare data is important for the medical experts to provide better services to their patients. Here is the discussion of the options and practices related to this process.

Information Sources: Medical imaging archives, picture archiving and communication systems (PACS), wearable devices, and other relevant sources of healthcare data should be identified by the medical experts (Pandey et al., 2022). Team up with medical care suppliers and establishments are required to get close enough to the essential information.

Security and privacy of data: Stick to severe security and security conventions to safeguard delicate patient data can be taken as a best practice. Keeping patient confidentiality by adhering to laws like the Health Insurance Portability and Accountability Act (HIPAA) is an important part of the collection of healthcare data.
Qualitative Data: Examine the collected data for accuracy and quality. To address any inconsistencies, missing values, or errors that could hinder the performance of the image processing algorithms, there is a need to employ data cleaning and preprocessing methods.

Storing

Image processing depends on healthcare data being stored effectively by taking into account the following options and best practices:

Online storage: Use of safe cloud storage options are taken by the medical experts to store healthcare data. Scalability, accessibility, and backup capabilities are provided by cloud platforms. The medical experts try to carry out encryption and access controls to safeguard the put away information (Jyotiyana and Kesswani, 2020).

Information Lake/Store: Creation of a centralized data lake or repository is required to consolidate healthcare data for image processing. This considers simple recovery, sharing, and joint effort among specialists and medical care experts.

Formats and Standards: Stick to standard configurations like Advanced Imaging and Correspondences in Medication (DICOM) for clinical pictures and Wellbeing Level 7 (HL7) for clinical information is helpful to store the medical data and use them properly in image processing. This guarantees similarity and interoperability across various frameworks and works with information sharing and reconciliation.

Sharing Medical Information for Image Processing

Sharing medical services information is significant for cooperative exploration and working on quiet consideration. Think about the accompanying prescribed procedures:

Agreements for the Sharing of Data: A proper layout of information sharing arrangements or agreements that frame the terms, conditions, and limitations for information sharing are followed by the medical experts to share the essential data appropriately (Tchito Tchapga et al., 2021). This guarantees lawful and moral consistency, safeguarding patient security and licensed innovation privileges.

Techniques for De-Identification: Patient-specific information can be anonymized from the shared data using de-identification techniques while still remaining useful for image processing. Data can be shared in this way while privacy is maintained.

Transfer data safely: Encrypted channels and secure channels for data transferring are very much required to transfer the healthcare data. It helps to maintain confidentiality and prevent unauthorized access or interception because it can harm the treatment process. Safe transfer of data also helps the medical experts to improve their services and get better responses from the patients.

Documenting

Healthcare data must be properly documented for long-term reproducibility and usability. Here is the discussion on some options and practices related to documentation of the healthcare data for image processing. Most of the time, medical experts are trying to catch and record thorough metadata related with medical services information, including patient socioeconomics, securing boundaries, and preprocessing steps. This data helps in grasping the unique circumstance and guaranteeing information discernibility (Willemink et al., 2020). Documentation of the healthcare data is very important and the medical experts try to do this in a proper way for providing better services to the patients.

Maintenance of the data

Version Management: Medical experts have tried to implement version control mechanisms to keep track of changes to the data, algorithms, or preprocessing methods over time. Reproducibility and comparison of results are made possible by this.

Governance of Data: Medical experts have tried to establish data governance policies and procedures to guarantee data integrity, accessibility, and compliance with regulatory requirements (Ahamed et al., 2023). They should check and update these policies on a regular basis to keep up with new technologies and best practices.

Healthcare data for image processing must be accessed, collected, stored, shared, documented, and maintained with careful consideration of privacy, security, data quality, interoperability, and compliance. Researchers and healthcare organizations can harness the power of healthcare data to advance medical imaging and patient care by adhering to best practices.

Propose a question that could be asked in relation to the unstructured data and what software might help to run AI and answer the question

Question: "How AI is used to improve lung cancer diagnosis accuracy by analyzing medical images from unstructured data?

Tensor Flow can be taken as software that might be helpful in running AI algorithms to answer this question. Tensor Flow is an open-source library broadly utilized for AI and profound learning undertakings, including picture handling. It gives an exhaustive system to building and preparing brain organizations, making it reasonable for creating computer based intelligence models to break down clinical pictures for cellular breakdown in the lungs location (Amalina et al., 2019). The extensive ecosystem and community support of Tensor Flow also make it possible to integrate other image processing libraries and tools, making it easier to create and implement accurate AI models for better healthcare diagnosis.

References