Make your own jarvis program12/31/2022 Last question, do you find that it's "slow"? I mean, by the time I speak and the speech processing engine (I'm also using Google Voice Recognition, so sending packets, processing, and responding take time) finishes, it then needs to make some more requests to actually fetch my information/do something, then maybe some more local processing, it could be 5 seconds before I get a response. Right now I wrote some plugins for it to get the weather and get stock ticker values for me, but those aren't really useful when I think of a personal AI.įor a while now I've been trying to get into Machine Learning, maybe this is a good opportunity to try to grab "context" from what I'm asking of it and perform more complex tasks ("I have an image on my desktop called test.jpg, can you resize it and email it to me student email address", but that's probably just a far off dream).Īlso, are you only using OpenCV to determine if you are sitting in front of your computer, or other gestures maybe to perform certain tasks? For the authentic Jarvis voice I recommend using IVONA as they have a cloud system. What kinds of automatic desktop tasks do you do? After reading this yesterday I just got a simple personal AI setup, but I'm struggling to find real problems it can solve. I am currently using Google voice recognition and so far it works the best. I'll look into starting a Youtube series soon and post the link here so keep your eyes peeled! And I will make sure to pm all the good people who have replied to this and shown interest. OpenCV is the hardest so ignoring that the rest can be set up within a few days. I can do a youtube guide if there is enough interest.Įdit: Quite a bit of interest here! As mentioned, yes there is a lot of different technologies working together and even more that I haven't mentioned so it will probably have to be a tutorial series if I get down to it. ![]() So I use my android phone's mic for voice detection and then send the queries to my laptop using web sockets or through GCM. This was of course very bad until one day I realised that in my pocket lies a phone with a very good mic for call quality. Hardware-wise my earlier iterations used laptop inbuilt mic when working with windows sdk. Another cool feature is that I have used Selenium to scrape data from Google and Wikipedia so if I ask it to define something it reads me a short description. It also has face recognition using OpenCV so when I am not in front of the desktop, the screen turns off automatically. ![]() I am currently using Google voice recognition and so far it works the best.įor the authentic 'Jarvis' voice I recommend using IVONA as they have a cloud system where you can send a text and they send you the voice mp3 which you can then play.įunctionality for my project ranges from notifying me about events on my calender to doing automatic tasks on my desktop. I have used windows speech recognition, sphinx recognition and then dragon speech recognition sdk and in all 3 cases the results were not good enough.
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |