Title：Gadetary -- Gadget Private Secretary
Displayed Name：FactoryIOT Research Team
| Concept / Overview |
|---|
| Have you ever forgotten that you have met an important business contact before, and later created an embarrassing negative impression? Private secretary "Gadetary" takes a photo of who you are meeting and stores that person's picture in association with a voice recording in his/her profile. When you meet him/her at a later date, Gadetary performs face recognition by camera, then reads the corresponding profile to you via an earphone |
Development challenges
Challenge 1: When we thought about how we can use GR-PEACH in our solution, our 1st challenge was how we can make the gadget fit nicely on our business outfits. We planned the gadget to look like a corsage or handkerchief.
Challenge 2: How to ensure the gadget takes a photo and operates naturally? We needed to test the gadget to make sure it works even with bad lighting and a moving subject.
Challenge 3: When performing face recognition, the challenge is how to match a picture in the database with the person in front of the camera. Since there are various recognition services available, some free and some for purchase, we plan to test and select from among them.
First, we will try the face recognition solution in GR-PEACH AUDIO CAMERA Shield.
Challenge 4: We don't know what programming issues occur related to face recognition in Skype or other lvideo lectures. We welcome your feedback about this issue.
Our first target application is for general business scenes. This device is for those who are not good at remembering faces, those who meet large numbers of people, and those afraid of making mistakes in communication .
As a finalist, if we do win and receive prize money, we will try to furhter miniaturize the dedicated board as well as enhance the design so that it can be worn in any social situation. We can eventually target this device not only for those who are not good at remembering faces and names, but also for who are uncomfortable at communcation in general. This gadget will help various people feel easier about talking to anybody and expand their sphere of colleagues and friends.
Demo video
Operation
As soon as GR-PEACH is powered on, the attached camera starts running.
GR-PEACH takes a photo when it finds a person.
Then GR-PEACH accesses the face recognition database in AWS. Once GR-PEACH finds the person registered in the database, GR-PEACH plays back the pre-recoded voice memo of the person from the Bluetooth earphone connected to GR-PEACH. (For example: "Gadget Renesas, Mr. Yamada. Exchanged business cards at the February exhibition.")
In this way, the user can remember names and profiles of anyone he/she has previously met.
If there is no registration of a certain face, GR-PEACH stores the picture and tells the use that the current person has not been registered yet.
The profile and voice memo can be registered on a dedicated smartphone site and data managed in AWS.
How to connect to AWS (AWS IoT)
https://aws.amazon.com/jp/iot-platform/?nc2=h_iot
Save image/Save audio (Amazon S4)
https://aws.amazon.com/us/s3/?nc2=h_m1
Creating a trigger by uploading (temporarily saving) the image (AWS Lambda)
https://aws.amazon.com/jp/lambda/?nc2=h_m1
Recognizing a person (Amazon Rekognition)
https://aws.amazon.com/jp/rekognition/?nc2=h_a1
Connecting people and voice (Amazon DynamoDB)
https://aws.amazon.com/jp/dynamodb/?nc2=h_m1
FactoryIOT Research Team
GR Design Contest Japan 2017 Finalist
