Xinyu Zhao:
What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).
- I tried the C930e from Logitech, the resolution is good and the boundaries of the objects after the edge capture, especially the table are much more definite than previous ones. However sometimes the it would fail to run with VideoCapture.read() returning False, giving the cpp:181: error. The other model from Genius had some temporary problem with its software, I also experimented more with ndarrays in numpy.
Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?
- It’s somewhat behind schedule, partly because of the midterms, after progress would be more normal.
What deliverables do you hope to complete in the next week?
- Fixing the problems encountered above.
- keep working on transferring data to the next module
Tianhan Hu:
This week I continue to work on hand-detection with dynamic threshold. My attempt was to get the average value of the first few frames, and contract that with the frame when hands are engaged in the picture. However, this method has several drawbacks such as it could not effective differentiate arm and hand. Also, the accuracy did not increase much with the new approach.
After discussing this issue with my teammates, I was suggested that I could potentially use machine learning for hand-detection and even for hand-gesture detection as a whole. I am starting to look into that this week, and might keep on researching in this direction in the following weeks.
0 Comments