DeepFake in Medical Science

X-rays can cause changes in our DNA and, as such, may provoke danger farther away. Hence, X-radiates have been named a malignancy causing specialist Trusted Source by both the World Health Organization (WHO) and the United States government. While X-radiates are associated with a barely extende

Project Title

DeepFake in Medical Science

Project Area of Specialization

Artificial Intelligence

Project Summary

X-rays can cause changes in our DNA and, as such, may provoke danger farther away. Hence, X-radiates have been named a malignancy causing specialist Trusted Source by both the World Health Organization (WHO) and the United States government. While X-radiates are associated with a barely extended risk of infection, there is an all right of transient outcomes. Receptiveness to high radiation levels can have an extent of effects, such as hurling, passing on, fainting, going uncovered, and the inadequacy of skin and hair. Other than this heap of results it is much more extreme also on the grounds that occasionally a solitary bio-picture isn't adequate or now and again it is hard to investigate. You need to see a case from various perspectives to improve understanding since it's anything but a question of human life. Include a period at the end of a run-in heading. Note that you can include consecutive paragraphs with their own headings, where appropriate.

Project Objectives

The main objective of our project is to bring advancement in the field of medical science specifically (bones), with help of our project there will be multiple datasets that can be helpful to visualize the disease, that affected bone significantly as there will be multiple the images generated with the deep fake technology. It can save time also as in multiple cases where an x-ray is required with different variations and it takes time while with the help of our project from a single data set or an image the variation can be provided in real-time and there is no need to wait in queue or to collect the reports, X-Rays when we can get results in realtime. Also, X-beams can cause transformations in our DNA and, in this manner, may prompt malignancy further down the road. Therefore, X-beams are named a cancer-causing agent Trusted Source by both the World Health Organization (WHO) and the United States government. While X-beams are connected to a marginally expanded danger of disease, there is a very okay of transient results. Whereas our project does not produce any kind of harmful rays and is environment friendly that can be useful for the future generation also Deep-Fake in Medical Science

Project Implementation Method

SYSTEM DEVELOPMENT PROCESS: We trained the huge data set of the X-Rays to our system with the help of GAN by which we trained our machine to detect the X-Ray who have a Bone on their hands so the system will work just like that to detect and catch them on the spot by generating an image of its multiple variations on run time to check the result.

SYSTEM ARCHITECTURE: We have designed a simple model of GAN for generating variations of bioimage from a single image. Using GAN, we will generate fake images from a real image for variation for this target we design a model. In our model, we train real images to gather from different and we trained them to make our dataset. The user puts an input image it will first serialize it is the process of converting an object into a stream of bytes to store the object or transmit it to memory, a database, or a file. Its main purpose is to save the state of an object in order to be able to recreate it when needed. Then we apply line spacing over that input. After this it will go to the generator for further creation of fake images, that will compare from the trained dataset, and the discriminator will decide whether the generated image is similar to the real one or if not then it will again be sent to the generator to generate better image this cycle will continue unless or till generator will generate the fake image similar to real one. For this achievement, we apply binary classification to get the condition of true or false if the generated image is similar then 0 means close to the real one otherwise false mean not close to the real one. We also apply Fine-tuning over the generator to get help from another pre-trained model and to make our model light weighted. We use a dataset of other image generating models by means of transfer learning. After this it will go to the generator for further creation of fake images, that will compare from the trained dataset and the discriminator will decide whether the generated image is similar to the real one or if not then it will again be sent to the generator to generate a better image this cycle will continue unless or till generator will generate the fake image similar to real one. For this achievement, we apply binary classification to get the condition of true or false if the generated image is similar then 0 means close to the real one otherwise false mean not close to the real one.

Benefits of the Project

With this project, we aim to bring the advancement in the field of Deep Fakes by creating the most accurate fake image with different and multiple variations compared to the Page 13 of 20actual images so that with several images as data it would be easy to analyze the problem in the bone. This project can help medical science to evolve with the most accurate results in no time as it will be helpful for future generations as well as in the current era.

Technical Details of Final Deliverable

The implementation phase of the project had brought up so many challenges in the beginning due to the designing, assembling of the data, research, and development of the project, and understanding of the literature to bring out main topics to focus more. The simulation of the project and assembling of the documents. Overall, it was a more involving and learning experience. Overall, the working of the project was quite made on time and then the testing phase of the project was done on time to make things easier. We had run our images through basic GANs and after getting results from this, we will show these results on the doctor portal where doctors can check results on large scale and got different variations in series. We deployed some python libraries and APIs to connect with Web Portal with Google Collabs. 

Final Deliverable of the Project

Software System

Core Industry

IT

Other Industries

Medical

Core Technology

Artificial Intelligence(AI)

Other Technologies

Artificial Intelligence(AI), Augmented & Virtual Reality, Clean Tech

Sustainable Development Goals

Good Health and Well-Being for People, Decent Work and Economic Growth, Responsible Consumption and Production

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com