We are looking for a software development company for this project who have expert developers skilled in both C# and Python. And the developer needs to be expert in machine learning too. We have tight deadline. So we are looking for a software development company or team, who have all the developers in house and can complete these tasks very quickly by assigning the tasks to multiple developers.
1. Let us know how many developers can you give us for this project. Let us know whether they are skilled in Python or C# or both? You will be asked to share some sample codes of those developers in Python and C# from relevant projects. And the developers will be given some codes in Python and C# from our end to rate the quality of codes.
2. Share your existing facial recognition project demo video if you have any.
1. Our software is for finding lost persons online using there photos. This is web based software hosted in server. We have different exe files for user software and admin software. Users and admin use the exe setup file to install the software in their PCs. All the software users are connected to our central server. We already have the software built in C# (.net) except the facial recognition feature. In this video, you can see how the software works and what are your tasks: [login to view URL]
Here's the text transcription of the video where the tasks are listed according to the video: [login to view URL]!AkZvryP4TXsfce3P5znto4eCpKY?e=eKVosT
2. We will send you our existing C# (.net) software as ZIP file. A number of your developers need to start working on the tasks shown in the above video. For facial recognition, you can use either Insightface or FaceNet Python library. Git links: [login to view URL] and [login to view URL]
If you use Insightface, you may use their model pre-trained in Glint360K dataset: [login to view URL]
Glint360K dataset contains 17M faces so we won't need manual training in any other dataset.
You may check Insightface Model Zoo page for more details: [login to view URL]
If you use FaceNet, you may use their model pre-trained in VGGFace2 dataset: [login to view URL]
VGGFace2 dataset contains 3.31M images, so we won't need to manually train the model in any other dataset either.
Whether you use Insightface or Facenet, you will need to use [login to view URL] to integrate that into our existing C# software.
So, I am giving you an alternative option too. The Github repo "Compreface" provides docker based API for Insightface and Facenet. You may call Compreface Python API from our C# codes for facial recognition purpose, if it's easier for you. You can use Glink360K pre-trained model of Insightface or VGGFace2 pre trained model of Facenet with Compreface. Here's the Compreface Git repo: [login to view URL]
Either way you go, we expect you to complete the tasks shown in the video within 3 working days maximum. Because we have a tight deadline.
3. Our server OS is Ubuntu. You will have to host the C# software on Ubuntu server OS possibly. If you can't host it on Ubuntu, you have to erase Ubuntu from server and install Windows instead if you need. We want you to install Webmin control panel in the server so that server management becomes easy for us. Once the software is hosted in the server, you will have to generate exe for user app and admin app separately. And you will need to complete all these tasks within the first 3 business days too. So, we require you to give us further developer/server admin for speeding up these tasks. You should also let us know how much server RAM, storage and CPU does the software need.