1.1 Problem Summary
There is this great problem of large amount of data being produced by medical apparatus which becomes too much to handle for a human. Or in some cases, there is la ack of specialist doctor needed to examine that data in order to diagnose a disease. Medical science with the use of information technology and in particular the use of machine learning can benefit from it.
Alzheimer being a neurodegenerative disease, it is hard to manage after certain stage. If in brain scan it is early diagnosed it can be helpful in greatly managing in coming time. But there is large amount of data and very few experts, and so there is the problem of detecting Alzheimer from brain scan images and that we hope to solve.
1.2 Aim and objectives of the project
Our aim is to provide a machine learning-based system that can precisely diagnose the Alzheimer brain patterns from the brain scan data it had been given. Then it can handle large amount of brain scan data and with the accuracy we strive to achieve, it can be really helpful for medical practitioners.
A web-based system which can be helpful in early detection of neuro-degenerative disorder is really helpful. Alzheimer being diagnosed without the help of neurologist and with the less effort and great accuracy is our primary and to help medical system.
1.3 Problem Specifications
A typical brain slice can not be distinguished from the normal brain slice. There are different parameters to check for in the different regions of the brain image. All of these things with the catalogued training data set that we provide are learnt by our model to diagnose Alzheimer from unknown dataset.
1.4 Brief literature review
We have done extensive web research and found many patents of which we have created various PSAR reports. Below image is sample of the search that we have done. Detailed PSAR reports can be found in Appendix.
We also have referred to various articles and papers which are mentioned in reference section.
1.4 Materials / Tools required
- High end GPU based system for to train our model. With the use of GPU we can train our model with images and process those images.
- Google colab
- VS Code
2.1 AEIOU Summary
In AEIOU summary we have following parts:
2.1 .1 Environment:
The surroundings and the nature of the place we visited.
we visited medical institutions where we found patients and we observed interactions among these patients.
The objects we observed during the visit are mentioned in below. The detailed list is shown in Figure 2.1.
We found activities like medicines being bought, treatment, diagnosis etc.
The users we observed were patients being treated and other users which are hospital staff. Further users are shown in figure 2.1.1.
Figure 2.1.1 Canvas of AEIOU
2.2 Empathy Mapping
Here, we have took a doctor Raj Patel as shown in figure.
The stakeholders are depicted in Figure 4.2.
Figure .2.2.1 Stakeholders
Figure 2.2.2 Story
Figure 2.2.3 Canvas of Empathy Maping
2.3 Ideation Canvas
In Design ideation: the conceptual sketch in the digital age Ben Johnson defined ideation as ‘a matter of generating, developing and communicating ideas.’ The various aspects of the Ideation canvas prepared by our team are shown in Figure 2.3.1- 2.3.5.
This section can have all the people involved and related to the domain of medicine.
Figure 2.3.1 People
This section has activities of all the people mentioned above in the “People” section in the ideation Canvas.
Figure 2.3.2 Activities
In this section, we need to find in which situation do they do.
Figure 2.3.3 Situation/Context /Locations
2.3.4 Props/Possible Solutions:
Props are the keywords which may flash in our mind by thinking of some issues which may occur by combining random people, activities and situation/context/location sections in the ideation canvas.
The props for our design project are shown in figure 2.3.4.
Figure 2.3.4 Probes/Tools
Figure 2.3.5 Ideation canvas
2.4 Product Development Canvas
Our aim is to provide diagnosis of Alzheimer using machine learning.
Figure 2.4.1 Canvas of Product Development
- · Our road map to project is very simple. It can be understood with the use of the following flow diagram. By which, it can be well understood how we are to implement and develop and then in the end deploy our product.
Figure 3.1 Alzheimer Project flow
- · This are some of the Alzheimer patients brain scan slices that we have trained our model with.
Figure 3.2 Brain Slices
- · This is the snippet of the Alzheimer machine learning model code that we have developed. It has accuracy of 70-75%.
Figure 3.3 Machine Learning model
This is the Next step is the pre- processing of the image of the brain slice into grey-scale that we want to check for the Alzheimer.
Figure 3.4 Image Pre-processing
- · We can see the given brain slice is of Alzheimer patient’s or not from our resultant output. This is how our model works with current training of model.
Figure 3.5 Final result
4.1 Advantages of our work:
4.2 Unique features of our work:
- · Our work makes the work of medical practitioners easy.
- · It is a unique way to handle and diagnose disease from large amount of data.
- · It provides a web-based interface for processing of a brain slice to know if it is Alzheimer or not which will be very helpful in rural areas.
4.2 Scope of future work
- · Our road map in future includes web-based User Interface which will be helpful in using out product.
- · Machine learning model code that we have developed. It has accuracy of 70-75%. Which we plan to improve to 95% or more with upcoming future work.
- Alvarez, I., Gorriz, J. M., Ramirez, J., Salas-Gonzalez, D., Lopez, M., Puntonet, C. G., et al. (2009a). Alzheimer’s diagnosis using eigenbrains and support vector machines. Electron. Lett. 45, 342–343. doi: 10.1049/el.2009.3415.
- Anagnostopoulos, C. N., Giannoukos, I., Spenger, C., Simmons, A., Mecocci, P., Soininen, H., et al. (2013). “Classification models for Alzheimer’s disease Detection,” in Engineering Applications of Neural Networks, Vol. 384(Pt II), eds L. Iliadis, H. Papadopoulos, and C. Jayne (Berlin; Heidelberg: Springer), 193–202. doi: 10.1007/978-3-642-41016-1_21.
- Chaves R, Ramirez J, Gorriz JM, Illan IA, Gomez-Rio M, Carnero C, Alzheimer’s Disease Neuroimaging Initiative. 2012. Effective diagnosis of Alzheimer’s disease by means of large margin-based methodology. BMC Medical Informatics and Decision Making 12:17.
- Dong Z, Liu A, Wang S, Ji G, Zhang Z, Yang J, Zhang Y. 2015a. Magnetic resonance brain image classification via stationary wavelet transform and generalized eigenvalue proximal support vector machine. Journal of Medical Imaging and Health Informatics 5:1-9
- Eliasova I, Anderkova L, Marecek R, Rektorova I. 2014. Non-invasive brain stimulation of the right inferior frontal gyrus may improve attention in early Alzheimer’s disease: a pilot study. Journal of the Neurological Sciences 346:318-322
- Eskildsen, S. F., Coupé, P., Fonov, V. S., Pruessner, J. C., and Collins, D. L. (2015). Structural imaging biomarkers of Alzheimer’s disease: predicting disease progression. Neurobiol. Aging 36(Suppl. 1), S23–S31. doi: 10.1016/j.neurobiolaging.2014.04.034
- Gomes, T. A. F., Prudêncio, R. B. C., Soares, C., Rossi, A. L. D., and Carvalho, A. (2012). Combining meta-learning and search techniques to select parameters for support vector machines. Neurocomputing 75, 3–13. doi: 10.1016/j.neucom.2011.07.005
- Kalbkhani, H., Shayesteh, M. G., and Zali-Vargahan, B. (2013). Robust algorithm for brain magnetic resonance image (MRI) classification based on GARCH variances series. Biomed. Signal Process. Control 8, 909–919. doi: 10.1016/j.bspc.2013.09.001
- Lopez, M., Ramirez, J., Gorriz, J. M., Alvarez, I., Salas-Gonzalez, D., Segovia, F., et al. (2009). “Automatic system for Alzheimer’s disease diagnosis using eigenbrains and bayesian classification rules,” Bio-Inspired Systems: Computational and Ambient Intelligence, Vol. 5517, eds J. Cabestany, A. Prieto, F. Sandoval, and J. M. Corchado (Berlin: Springer-Verlag Berlin), 949–956.
- Savio, A., and Grana, M. (2013). Deformation based feature selection for computer aided diagnosis of Alzheimer’s Disease. Expert Syst. Appl. 40, 1619–1628. doi: 10.1016/j.eswa.2012.09.009
- Streitburger, D. P., Möller, H. E., Tittgemeyer, M., Hund-Georgiadis, M., Schroeter, M. L., and Mueller, K. (2012). Investigating structural brain changes of dehydration using voxel-based morphometry. PLoS ONE 7:e44195. doi: 10.1371/journal.pone.0044195.