We are looking for a experienced developer in Computer Vision/Machine Learning.
We intend to develop a tool that uses Computer Vision/Machine Learning to identify if a photo contains a Swiss Identification Card (Swiss ID).
The tool has to work on color photos taken from various mobile phone devices (iOS an Android). The tool returns a confidence/score if the photo contains a Swiss ID or not.
The target application is to sort out photos with Swiss ID’s versus photos without Swiss ID’s.
- The developed tool must run on Windows OS.
- Prefered language is C#, however the developer can suggest other languages he/she might be more familiar with
- Use open source/easy available tools such as OpenCV, Matlab etc.
- Focus will be on approaches that use Computer Vision and Machine learning.
The project will be in two phases.
Phase 1: Prototype and Proof of Concept
The developer will provide a compiled version of a simple prototype that demonstrates the capabilities.
The prototype must run on windows, but might be a simle WinForm, Command line tool or other simple way to execute the program.
The prototype must accept as input a training set. The training set is one or more real Swiss ID’s which represents a good sample of Swiss ID’s.
The prototype must be able to process a number of documents in a specific windows folder and output the result as a text file. The output will be the file name and the confidence/score with comma separation.
The prototype/proof of concept is deemed a success, if the test documents are correctly identified as Swiss ID/non-Swiss ID photos.
Phase 2: Final Component
If Phase 1 is a success, Phase 2 can be awarded. Phase 2 is the implementation of the prototype technology into a more stable component.
Specifically, the component needs to have an API which can be called with the necessary parameters.
The Component must be able to store the settings/learned samples.
If you are interested, we are happy to discuss in more details on the scope/challenges and targets of the projects. We are also looking forward to have an effort estimation for both phase 1 and 2 before the project is awarded.
The attached document contains some additional information, especially regarding sample documents.