Hiring a programmer to continue working on an existing computer vision project in C++ using openCV and several other tools. The aim of the project is to detect the video and analyze developer’s behavior. The program will extract the that window frame from a recorded video of user's action which contains Eclipse IDE frame, when the user encounters syntax error, the program should provide relevant Google search results to help user address the syntax error.
The current treatments include(may have bugs) the following steps.
1. Step 1: convert video to frames. In our existing framework, the frames of videos are taken as input data. The converted tool is named ‘ffmpeg’ which is an open source tool.
2. Step 2: Difference analysis. There are many frames when convert the original video to frames. The aim of the step is to reduce processing frames. The attached M file is used to achieve the goal.
3. Step 3: Detect contents. The application object name and order of the frames are detected in this step. The results are recorded in database. The project is based on the development of VC2010 and OpenCV-2.4.3.
4. Step 4: recognize the contents . Especially, the top object is cropped and saved as images in step 3. The images are recognize by an Optical character Recognition software named FineReader. The results are also saved in database.
5. Data Analysis (This is to be done!)
Requirements/Deliverables:
1) Real-time processing
- Performance should be fast; current program is taking some time to analyze
2) Feedback to user after analysis
- After detection, the program should provide relevant Google search results to help user address the syntax error
3) Formal Technical Report
- Guild lines will be sent separately.