Page 32 - Veritas

FORENSIC SCIENCE GLOBAL NEWS UPDATES

that deals with the processing of evidence (i.e. from the crime scene to the
lab and from the lab to the court of law) for justice. Digital forensics is the

branch of Forensic Science that deals with the analysis of digital evidence.
It consists of identification, acquisition, preservation, transport and analysis

of digital evidence.

The success of AI is data-driven and there is no specific code or program-
ming that controls the output so far. In computer science, AI is split into

two- Machine learning and Deep learning. In machine learning, the fea-
tures are designed by human engineers, unlike deep learning where fea-

tures are learned from the data using general-purpose learning procedures.
For the processing of information and evidence analysis in Forensic Sci-

ence, this general intelligence of deep learning would be significant. With
the advancement of Modus Operandi (MO) of criminals, it has become the

need of the hour to take help of such technological and scientific methods
for the purpose of investigation of crime.

Pattern recognition and differentiation in objects through experience are

among the many tasks performed by the human brain. Machine learning,
which is the application of AI, mimics this ability of humans and enables

them to learn from experiences. Facial recognition, fingerprint matching
etc. are the application of AI for the purpose of screening and identifica-

tion of the required data, from the huge amount of data for the purpose of
the criminal justice system.

Applications of AI in FS

1. The machine-generated data will increase the objectivity in the results

of the analysis.
2. It will lead to the automation of evidence analysis by the automated

reasoning method. It will reduce the overlooking of possibilities that may
be caused by humans.

3. There will be a reduction in variations in reports due to subjectivity and
different interpretations due to different experts.

4. It has the potential to compare the low quality/degraded images as well,
by lowering the quality of “standard” to the same extent as “questioned”

to get the match.

VERITAS 17