Difference between revisions of "PersonalLifeRecorder"
m (added links and such)
|Line 44:||Line 44:|
[[Category:]] <!--MAKE AS MANY CATEGORIES AS YOU NEED-->
Revision as of 14:25, 6 June 2012
Imagine a small device that you wear on a necklace that takes photos every few seconds of whatever is around you, and records sound all day long. It has GPS and the ability to wirelessly upload the data to the cloud, where everything is date/time and geo stamped and the sound files are automatically transcribed and indexed. Photos of people, of course, would be automatically identified and tagged as well.
Imagine an entire lifetime recorded and searchable. Imagine if you could scroll and search through the lives of your ancestors.
- Phase 1
Simple audio recorder via earbuds and mic on Android phone. Records all audio during the day. Press of a button on the app will dump the audio to a file on a network PC drive. Desktop app will then analyse the audio first transcribing into text. Text is then saved to a file and/or emailed to the user for later searching etc.
- Phase 2
Add more intelligence to desktop app so that after transcribing into text it will search for key words like "remind me to" or "remember this person" or just plain "remember this". This will then do things like add appointments to your google calendar, add a contact to your address book, or just send an email to yourself with a note. Make sure to create this part so it's easy for others to add key words and how the app will handle them.
- Phase 3
Add an automatic dump feature that senses the "home" wifi SSID and looks for the network drive previously set up. If found it will automatically dump the day's audio to the home PC and begin analysis without any input from the user. Perhaps adding a setting/option to auto dump at home arrival with or without a voice prompt "I see you're home now. Would you like me to save today's audio now?".
- Phase 4
Add a keyword for "go privacy mode" which tells the desktop app to ignore all audio from that point on until you say something like "go public mode" or similar.
- Phase 5
Add the ability to upload the day's audio remotely from a strange Wifi or cellular data connection.
- Phase 6
Add the ability to take pictures of people and things and associate them to the recorded audio. This way you can take a picture of someone and say the "remember this person" keyword followed by their name and number and address etc. and it will use the date/time stamp of the photo to associate to that person's contact info in your address book.
- Phase N
Add GPS info that will associate with the audio keywords and perhaps track your daily movements. Could be used to remember/add places to your address book. Add any other option I haven't thought of yet!
Links and such
Java speech to text http://cmusphinx.sourceforge.net/sphinx4/
Voxforge used by Sphinx http://en.wikipedia.org/wiki/VoxForge