Page 211 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications

Chapter 7 Early detection and diagnosis using deep learning 201

investigation of that information into clinical understandings that
help doctors in preparation and then providing attention, eventu-
ally resulting into healthier consequences, inferior expenses of
maintenance, and amplified patient fulfilment.
ML and DL [10] together are leading the headlines of the news-
paper nowadays; big company like Google has developed a ML al-
gorithm which is used in the identification of cancerous tumors
on mammograms [9]. Also, one of the top universities of the world
used DL algorithm to develop a model for identification of skin
cancer cells in the body. An article read in recent years described
the consequences of a deep machine learning algorithm, which
was intelligent enough to analyze diabetic retinopathy in retinal
images [9]. It is clear that ML places additional arrow in the quiver
[9] of clinical pronouncements.
ML advances the aforementioned to some progressions that
are improved than the others. Algorithms can deliver instanta-
neous advantage to sections with progressions that are not
only reproducible but are also consistent. There can be a vast
set of image data sets, which can be based on radiology, pathol-
ogy, and cardiology. The capability of ML models is very high,
i.e., they can be skilled to detect and understand the images,
with the power to classify irregularities in the results, and high-
light the sections that essentially require consideration, thereby
increasing the effectiveness of the progressions that are under-
taken by the model. Using ML for a long time for the patients
of one family will help the doctor to examine the specific
member of the family with the problems that occurred in his
or her family history. ML can bargain an impartial estimation
to progress competence, consistency, and exactness.
Medical practitioners make use of an exclusive stage to
examine information and coil it back in actual time to doctors
to help in clinical conclusions. At the time when the doctor
examines the patient, he or she enters the symptoms, all the
related data, the test results in EMR [9], and all the steps taken
by the doctor in aforementioned scenario that holds the ML at
its backend, which is analyzing all the symptoms, test reports,
and data provided by the doctor and helps him in taking better
decisions while diagnosing or suggesting a test. On seeing this
in long term, not only doctors get improved insights of how to
deal with the patient in their case, but also pharmacists can
check if the medicines prepared are worth selling or have more
problems in them, that is, if just in case the developed medicine
is not able to achieve the accuracy that has to be attained, then