Hi @yerayefalcon
Let’s discuss my use case first- Recently i got a contract from our military college, they need to setup AI labs with ai modules like- Face recognition, object detection, radar clutter detection, Aerial object detection, NLP base question/answering system and a few more.
I was searching for deployable solution with user interface and I found none. All of them are command/code based solutions with no user interface. After a lots of work we finalised the face recognition, object detection and NLP modules but got stuck on radar clutter detection and aerial object detection.
Later we decided to use Matlab for the radar clutter detection but we still have no solution for the aerial object detection. One day I found a post on Linkedin about the cvedia and downloaded it instantly but I was trying to run it on a virtual window instance but it was not working (you haven’t mentioned the Graphics card requirements on download page).
But today I managed to run the application on a system with GPU, I really like everything and the drone detection module giving us hope of building a aerial object detection module (it should be able to identify the different kind of flying object wether it’s a military aircraft, civilian aircraft, make and manufacturers etc. etc)
It would be great to have tutorials on training custom data sets and using it in cvedia, a short video would be very helpful. and I know it’s too much too ask for and it’s hard to implant but having a build-in feature for training custom datasets will take this application to the new heights.
Having a tutorial on training custom data, will bring lot’s of peoples to try, train and deploy and maybe you can build a community around it where users can share the deployable projects on cvedia.
Trigger’s should play an alarm or sound when someone detected inside a zone, so it can be used in any industry instantly.
User interface requires lot’s of fine tuning and changes so new users can adapt it (like any other modern application). Would love to colloborate on this, I have experience in UI/UX.
Unable to see the text on video (example- people counting, or vehicle counting), it would be great if you can make the text a bit bigger or add background layer behind them
A page describing most common error and their solution, Like today when I tried adding a custom drone video of detection, it was not working and returning following error- Onlnit function failed. Could not start Instance
@Rod (I’m eagerly waiting for your response and any help would be highly appreciated)
Nice to have feedbacks on clicks, like- when I double click on screeenshot it should give a feedback that the image is saved in following directory. As during the use i thought screenshot save feature is not working later I find out that it’s saving the image inside the cvedia directory only and i was searching them inside the download folder.
and I have lot’s of ideas regarding interface and ai application. cvedia website also requires design changes, as the current design system is very basic and missing the modern touch and doesn’t give the clear perspective of possibilities and cvedia-rt should be featured on the homepage.
One thing I know for sure that this application will change the way how individual/startups/corporate use and deploy ai. And building a community for sharing trained models/projects will give it instant growth and scale and will make it the next and first github/dribble/behance for AI.