May 28, 2018

Reflection: Challenges and Successes

May blog




I to I (2.0) signature presentation can be found at: https://docs.google.com/presentation/d/1aLC1swxWnzyosxshGG2-uG36VUWa7dy7yINtPANeugg/edit?usp=sharing

It was a big success on Friday, May 25th! I really appreciate the attention, passion, and support given by my audience in the room that morning.

In the reflection section of my presentation, I mentioned that "the one-year journey of signature has been full of challenges and eventually some successes." 
Challenges:
-- First of all, the scope of this project is quite broad to be carried on by a student 
-- Technology limitations blocked my project to some extent and made me frustrated when trying to build a prototype.
-- Time management, it was hard to balance my time between signature and my other tasks, especially during college application season.

Successes:
-- Glad I touched base on the deep learning technology world through this project
-- My I to I design and the innovative back-side filming concept were confirmed
-- Learned the startup business process and learned the real need of sign language translation


Before the D-day, I organized my complete presentation and ran through several times to practice, each time getting better and better.
My first draft of presentation was rough since I put in too much information but didn't pick only the most important ones and organize them to be clear.  Gradually I learned how to filter the minor information and focus on the major ones. The second draft was still beyond time limit, so I knew I had to write a better script and be concise. Lastly, I was able to practice with a finalized script and be confident knowing my material.

On Friday morning, I gave my presentation with calmness. The Q&A part really surprised me. A lot of people expressed their strong interest in my I to I product.
One said that her best friend is a deaf person, so she could totally relate to this real-world problem and really thought my project and effort would be helpful.
Some others brought up a new viewpoint: I really should work on the voice translation into text to begin with, and they for sure will want to buy it! The current problem is that for people have hearing loss, the hearing aids are expensive do not work very well, and it's really bad that people can't wear glasses conveniently if they wear hearing aids behind their ears. I to I can also function as a translation device allowing people to understand others' talking and signs on the street in different languages.
In this way, I to I as smart glasses seem like a perfect solution. I was really inspired by this perspective and their passion, so I want to look into voice recognition as my next step.

It's been a hard but great year working on my signature! I have learned a lot!

- End -

Presentation 2

April #2

In my signature presentation, I'm talking about my business model development and business plan writing as the second major part of my project.

I'm including the contents of a business plan, introducing the helpful template with instructions that I found online, and presenting what I've learned.

I also planed include the market size, marketing strategies, competitor comparison, and financial analysis work in the slides. However, I soon realized that these work done in both last year and this year are kind of too detailed to be talked through. To fit the lime length, I decided to leave out the market size to put labels and animations on the competitor and financial charts to make it simpler to look at.

Looking forward to finishing my presentation and get ready for May.


Presentation 1

April #1

Getting to the end of this year, I've started to prepare for my final signature presentation. First, I'm working on the technology part.

I talked to Dr. Zhao, my engineering mentor, about updates on my tech research and what are some future plans. We figured out that I should talk about the process of us looking at image recognition at the beginning and then switching to video recognition and other methods. (the process of finding the possible technology was challenging.) I'm developing my presentation based on this outline. I'm also going to include my innovative idea of back-side filming, with the explanation of why it is important.

I also talked to Mr. Liang, an voice recognition engineer at Microsoft and got his advise that I should consider starting with voice recognition in the real development of the translation system since voice recognition technology is pretty mature nowadays.

Hope I can finish my presentation smoothly as planned.


April 13, 2018

Business Plan Writing

March #2

In March, I also researched on the producer sections of the business plan. I learned about what software companies I can possibly contact and cooperate for research and development.

Business Plan Writing

March #1

I have been continuing working on my business plan writing. I went back to check my marketing, customer persona, etc. information and materials from last year's business plan presentation.
With that, I did more research on the market because of the product change from smart glasses product to the translation system software.

I also researched about distribution channel since the software needs to be provided on certain platforms.

For the second half of the plan document, I have been working on the outline.

February 28, 2018

Business Plan Writing

February #2

Business Plan Writing

Over February, I have been working on drafting the business plan for my product. Based on the resources that I found together with my mentor, I was able to follow the plan model and instructions to learn about the formal writing.

Since I have also been working on my Launch X business plan competition, I strengthened my knowledge and ability of data research and presentation. It corresponded with my signature plan document writing pretty well.
So I'm starting off the writing process with objectives and problem statement now. I will continue working on this in March and April.

More Tech Research

February #1

More tech research

This month, I did more research on sign language recognition technology. With my engineering mentor, I found various super helpful studies/journal articles investigating this topic. Some reveal the same concerns that I had: ambiguity of hand movements, poor performance of normal cameras, and ineffective computing model However, some others confirm my ideas: recognition nowadays can be achieved by deep analysis camera like Kinect and then deep learning neural network. Kinect camera will work to track movements, and deep learning network will then recognize the meaning.
I aim to follow this path and further discuss with my mentor to see how I can do a software simulation practically.
Also, I have got replies from Kitware company for further connection to its technical teams.



January 30, 2018

Recognition Method and Kitware Contact

January #2


Another progress that I made in this month is contacting Kitware company. As Mr. Calos and Ms. Bernard both have personal contacts at Kitware, I decided to draft an email describing my product and asking for advice on my current technical problem.
First of all, I organized my thoughts on the parts of my project overall, including the concept, the form of product, functions, and applications of deep learning technologies.

Then I focused on the method of sign language movement detection (sign language sample collection is where the technical limitation is at). In terms of sign language recognition, the traditional way is that camera is held by a hearing person who is opposite to the hearing impacted person and the camera lens is pointed towards to face of the person who uses sign language, but I proposed a new method that I think would be more user-friendly for the sign language users—The camera installed in the smart glasses will capture hand movements with the lens pointed away from the body through a 180-degree peripheral vision, in order to make the product more convenient for hearing-impaired people.

I explained my ideas and concerns in my email to Kitware. I look forward to hearing back from them and hope to make further connections with their technical teams for further consulting.