The other day I was watching Sundar Pichai's keynote during
the I/O 2017, where he introduced Google Lens, an intelligent image computing
product from Google. The first thought that came to my mind was Arnold Schwarzenegger
from Terminator 2- Judgement Day, were the T-800, a cybernetic organism from
the future looks at an object and the details of the object pops up in its
vision. I imagine a day, when this technology gets more mature and the VR head
gears become lighter and cooler perhaps like eye glasses, and we can all be
like the T-800, where we look at people and their online profile pops up. Maybe
it could tell me that the tall guy in front me in the queue is my 5th
grade classmate...
Then I started thinking about what this could mean in the
ERP world. We already see benefits of OCR in processing invoices with zero
touch, etc. Some image processors are able to automatically process
applications and detect photos where the applicant's eyes are closed or identify
that the photo is not as per specification. But what Google is talking about is
altogether at a different level, with machine learning capabilities. Having
used some of the Google API's to generate QR code and to find geocode, I expect
relatively easy to learn & use APIs from them, that would abstract the
underlying complexities and return the most appropriate result for our request.
My thoughts on few of its applications in the near future...
·
Inventory Counting - Cameras can be kept along
the warehouse so all items are visible. We should be able to count each item
and their location in a blink.
·
Login - May be this will give us a robust facial
authentication mechanism to login!
·
Attendance - Place cameras at vantage points and
the in-time and out-time of employees can automatically be recorded. I don't
need to keep wiping my hands because it is wet and the biometric readers think
I am an alien!
·
Field Service - Camera could watch the field
service technician fixing let us say an electronic equipment and replacing a
circuit board. The parts used can be re-ordered and debriefing recorded automatically.
·
Expense -
System already has the field service technician's task and schedule. Let us say
the technician is on a two-day visit to perform a planned preventive
maintenance of a rig. On showing the lunch receipt to the mobile camera, it understands
it is a food receipt, it picks the date in the receipt and understands it is within
the visit's date range, gets the amount and records the expenses against the
task.
·
Inspection - As part of inspection during material
receiving, determine if the product meets the color specifications, if there
are any scratches, cracks, defects etc. and automatically receive or reject
material and record it in the application.
·
Enterprise asset preventive maintenance - Today
we have sensors monitoring enterprise assets and raising preventive maintenance
service requests. This can become better for e.g we can have photos taken of
the asset at intervals and understand changes like color of a tube, bulging of
the pump, rusting of the bend and raise preventive maintenance service requests.
Being an Ironman fan, I love the image processor part of
Jarvis and the glasses Tony Stark wears in Iron Man 3 that does image computing
and also projects images. I expect that in the next few years something similar
will be so commonplace that not just Mr. Stark but all of us can experience it!