datafly, biometric identity and document proof

13
Datafly, A Proposal for Mini Research Project Pretext Data collection, authentication and digitization is single most problem for various sectors like mobile sector, banking, loan and so on. But even though data can be used in a more generalized contest, we would restrict our discussion to Telecom industry. The case study will reflect our aim at offering solution for Telecom industry. In India, Telecom is one industry which is probably cheapest. A pure water bottle is $.5. But you can get a prepaid mobile connection at 1/10th of a dollar. Yes that is right. You can get a prepaid mobile connection in India for Rs 5. You will be surprised to know that it is cheaper than a cup of tea in most part of our country. You go to villages, you will see only 6 hours of electricity a day, poor roads and infrastructure, but you will see virtually everyone with a mobile. From milkman to farmer to person riding carts every one uses mobiles. Tariffs are not very high either. But what does that mean of Operators or Telecom Service providers? It means acquiring customers. India is a ocean of people. So you always have new customers. You only have to have resources to get them. Then what is the problem? Well that is the problem. There are many service providers. Beside that in India, you can any time change your operator without changing number at any instance. So operators need to be on the edge and a step ahead of their competitors. So they recruit salesman who are referred as "Sales Executives " against American notion of "Executives" which definitely point to someone at the helm of business or his departments. So the executives take up their vehicle. (If you ever come to India, you

Upload: rupam-das

Post on 03-May-2017

221 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Datafly, Biometric identity and document proof

Datafly, A Proposal for Mini Research Project

PretextData collection, authentication and digitization is single most problem for various sectors like mobile sector, banking, loan and so on. But even though data can be used in a more generalized contest, we would restrict our discussion to Telecom industry. The case study will reflect our aim at offering solution for Telecom industry.

In India, Telecom is one industry which is probably cheapest. A pure water bottle is $.5. But you can get a prepaid mobile connection at 1/10th of a dollar. Yes that is right. You can get a prepaid mobile connection in India for Rs 5. You will be surprised to know that it is cheaper than a cup of tea in most part of our country. You go to villages, you will see only 6 hours of electricity a day, poor roads and infrastructure, but you will see virtually everyone with a mobile. From milkman to farmer to person riding carts every one uses mobiles. Tariffs are not very high either. But what does that mean of Operators or Telecom Service providers? It means acquiring customers. India is a ocean of people. So you always have new customers. You only have to have resources to get them.

Then what is the problem? Well that is the problem. There are many service providers. Beside that in India, you can any time change your operator without changing number at any instance. So operators need to be on the edge and a step ahead of their competitors. So they recruit salesman who are referred as "Sales Executives " against American notion of "Executives" which definitely point to someone at the helm of business or his departments. So the executives take up their vehicle. (If you ever come to India, you will be surprised to see that beside two eyes, hands and legs almost everybody has an engine and two tyres assembled in a single assembly called a bike-100Cc version) and some of them are less costlier than iPhone. That is right. You get these 'good' bikes for $5550 which gives good mileage. So you can ride 70 km for 1 Lt Petrol ( that right now cost us about $1.2). So these executives goes to villages, meets some shops there and give customer forms. customers will fill up the forms. Along with form data, an ID proof like Voter ID card, or driving license is copied ( In India the term that everyone understand is XEROX. So if ever you are in Indian streets needing a document to be copied, and don't find any shop with Copier services, just look around for XEROX). Also the document needs to be supplemented with an address proof which could be recent electricity bill, or phone bill or house tax receipt and so on. These documents are countersigned by the customer. Also, customer needs to put a recent photograph on the form.

The executive revisits the shop in two days time, collects the documents and passes on to verification department for processing. Each of these documents are manually processed and SIM cards are activated accordingly.

Page 2: Datafly, Biometric identity and document proof

Verification includes checking

a) Whether signature matches with that of signature in any of the documents

b) photograph is close enough to the photo in document

c) address given is same as address filled in the form

and d) if the age of the person is above 18.

So you can understand that not only does the company need to provide allowance for the executives, also a lot of time is spent in acquiring and validating the data.

Problem Data gathering in Mobile  and Telecom sector is a major constraint.  The executives needs to visit the customers, convince them for a switch or opt in, gather relevant document, bring that to the main office where it is sent to validation department and upon validation goes for activation. This conventional process has been one of the major headache for the Telecom players back here in India. Then you have other problem which includes missing document, inappropriate ones. And you do not also want to miss out the customers. Therefore there needs to be a good automation of this whole process. This is not only the problem of telecom sector but also many industries like loan sector are suffering from the same problem.    

Page 3: Datafly, Biometric identity and document proof

Current process of data gathering.  This data mainly contain three essential parts:a) Form and declarationb) Supporting Documents: Identity proof and address proofc) Photograph of the customer.Mostly the documents are copied version of the original countersigned by the customer. The process diagram will clearly explain you the problem. One Working Day's delay is due to the fact that once an executive gathers documents, he can not go back to office to immediately hand over the data. He needs to complete all the calls and even attend on the fly calls, before he can actually produce the documents for approval. Validation team generally checks whether the the photo of ID card and

Page 4: Datafly, Biometric identity and document proof

given photograph matches, whether  the copied document is clear, and if there exists any ambiguity in the document. If there are problems with documents, it is sent back to the main office and from there the executive who needs to recollect the data suggested by the validation department. This has been a major issue for the Telecom sector. At an age where everybody wants the things to be rocket speed and have little patience to bear, such delays often causes the companies with potential customers. Companies are trying to come up with better solutions but have failed. One of the primary reason for such failures were the absence of robust system.  Executive goes to consumer's places and the only medium of data collecting is either smart phone or tablets. Scroll through Apple store, Google play store, AppUp store and try to locate one good app that meets such real time business problems and you will find none.  It is partly because these devices are conceived more as entertainment devices and less business devices. Low processing capabilities are other major problems that have effected such solutions. Such business solution needs real industry experts for execution. Indie developers have been found to be more inclined to game and entertainment niche and major service providers have yet not being able to conceive the idea of tablet based solutions. Other major factor that has affected is Windows. It has been such a user friendly operating system over the years and supported such wide range of software and platform that many companies are still continuing with windows XP and sadly even Win98. Many of the retail solutions for small retailers are developed by small IT companies who are specialized in SME.

Page 5: Datafly, Biometric identity and document proof

      The hole problem now gets summarized to following issues:a) Lack of digitization in document and data gatheringb) Lack of developers focus and  interest and hence lack of Appsc) Cost deduction in IT infrastructures by many companies to meet global economic slowdown. While visiting Local Airtel ( Leading Telecom Provider in India) Office for discussing about the problem, they were more than happy, in fact overwhelmed to discuss the issue. " We get more than three thousand forms daily. More than 5% of them are faulty documents.  That leads to a very tedious process and puts immense pressure on all authorities. It would be an immense help if you can automate part of the process."    Was the quote from Mr. Babu, the manager in charge.   With the tablets supporting desktop mode, it is more conventional device with touch, voice, camera features, little extra RAM and processing capabilities to do little extra stuff. Therefore now, the 'desktop like' applications can be developed and ported to these devices with Sync feature to manage the data in either a server, cloud or local machine.  

Page 6: Datafly, Biometric identity and document proof

What the App does and How it Solves the Problem  At Integrated Ideas, we have really long history of working with SME. out products includes TraderPlus ( a software that is developed purely for distributors and have sold over 2000 licenses over last couple of years), Car Service Plus (http://www.appup.com/app-details/car-service-plus, which has sold more than you would prefer to agree.) Our other products include Police Admin Pro, the most comprehensive police department administration software deployed in many S.P. Offices across Karnataka,  CSPlus ( A complete package for computer sales and services) and many others.Off late I have been working on a project called Mobile Plus to automate the documentation of new connections and managing them more appropriately. 

Page 7: Datafly, Biometric identity and document proof

The above diagram very much explains what documents are collected from the customer. Interestingly Identity proof like Driving License always have photo, income proof could be tax document or bank statement and may not have a photograph. Address proofs are current electricity bill, or telephone bill ( postpaid only).  This is universal to almost any sector. Prepaid mobiles on the other hand has done away with address proof and any photo identity proof is sufficient. Copied ( XEROXED   " />  )documents needs users counter sign, which is cross verified with the signature at the bottom of the form.     So, what is so cool about Datafly and why is it claimed as a generalized solution even though it is very much industry specific at this moment? 

Page 8: Datafly, Biometric identity and document proof

Technical Overview    The executive acquires the details and first fed them into the form. The process is much easy with Lenovo tablets as a keyboard can be used flawlessly. Once the form elements are processed, the app asks the executive to take the photographs of the documents. The capturing is performed using EmguCV with C#. Once the document is captured, the software searches for a face in the document using EmguCV's face detection library. If faces are not found, automatically the document is rejected and system requests the executive to recapture the document or provide another document with clearer face. Once face is located, it is saved as reference face. It asks the executive to take a photograph of the customer. using the same face detection library the face part is segmented and snap of only face is taken by the system. This is matched with the reference face. Remember scale of both the photographs will be different. Hence conventional PCA based based face recognition will not work in this case. We need to adopt and implement Adaptive local binary pattern based face recognition system. Manhattan distance  of the normalized faces are obtained and threshold to check the percentage of match. If the percentage is high, the process authenticates the face. It follows that up by extracting the address from address proof using EmguCV's OCR Library.  Point to be noted is that as the address proof document is essentially a bill, it will have several text other than the address. TF-IDF based text matching will be adopted to match OCR text and address entered in the adress box. If validated, it provides the customer with an option to sign on tablet. He can use fingers or stylus to sign the document. Directional vectors from the signature can further be used for future verification.  It accepts the application and serialize entire text and images into a single xml document.  

Page 9: Datafly, Biometric identity and document proof

 Interesting part is that images can be converted to utf text using encoding. Thus photo, and images of proofs along with form elements can all be put in a single xml file. The executive need not to create any directory or follow any manual process. As xml is understood by all platforms including Android and iOS, developing the solution to a cross platform solution also becomes viable. One of the strongest argument that may come here is why not use HTML5 which is readily cross platform. OpenCV is not ported to HTML as flawlessly as with java and c#. As the whole App will be using extensive image processing techniques, I would rely on proven C# rather than emphasizing on portability.The xml file can be sent to respective authority either through a webservice or could be uploaded to cloud account or could even be emailed to the respective authority. For this app we are going to use SOAP and webservices with our own infrastructure and refrain from using a cloud.This decision allows us to design the solution so that data can be uploaded directly to Telecom provider's server from where it will be polled by validation agency. The solution is presented with the image bellow. 

Page 10: Datafly, Biometric identity and document proof

 

 Features of Tablet Used:  1. Front/Rare Camera2. Touch/Stylous 3. Better processing capability 4. Longer battery backup 5. Connectivity  

Target Users

Page 11: Datafly, Biometric identity and document proof

Datafly intends to simplify the process of collecting document from consumer/customer and validating. It is one of the most challenging business problems in today's validation driven businesses like Telecom and Loan Sector. Therefore the App is suited for any document collecting agency or business peers, like banks, hotels etc. Why generalized? If you look at the form, this is the format adopted by most businesses. Thus we claim that Datafly is suited for all industries. However no business solution can be proposed if not developed a base sector in mind. This is because the need will largely vary between the sectors. Hence Telecom industry and mobile customer's case study is adopted for the proposed app.