Home - UCI-EECS113-Spring17/Chris-Chan-Stealing-Your-Waifu GitHub Wiki

Welcome to the Chris-Chan-Stealing-Your-Waifu wiki!

Team Chris Chan Stealing Your Waifu

  • Paolo Caraos (ID: 54702118)
  • Christopher Chan (ID: 81425568)
  • Thomas Lee (ID: 27494035)
  • Kevin Nguyen (ID: 65512368)

Project Description and Features

For our object, we plan to create a program that removes objects from photos. We will do this by using the PYNQ board and a webcam to record a brief time-lapse video. With enough frames, the pictures can be compared and objects that have moved in the background can be removed by using object detection and movement detection algorithms. This would be useful in removing people walking in the background of a picture.

Five-Week Plan

  • Week 5: Submit proposals (written and video)
  • Week 6: Research on algorithms, video input, and frame separation. Create shared test video
  • Week 7: Implementation of algorithms and successful video input
  • Week 8: Debugging and testing
  • Week 9: Output testing and re-calibration
  • Week 10: Presentation/Documentation

Progress

As of now, we have tested a webcam while it is connected to the PYNQ board. We observed that the PYNQ board cannot take in too many frames per second, so we believe that we will likely have to utilize multithreading to take advantage of the PYNQ board's cores. Furthermore, we will also try to incorporate the MicroBlaze processors to assist with computation.

Required Sensors/Actuators

  • Camera (May be used to test the initial video input)
  • HDMI Output
  • USB 2.0 Webcam with 6-LED Illumination (Driver Free) Details

Task Breakdown

  • Paolo - Additional Image Processing, gray scale, face detection and overlaying accessory images
  • Chris - Gaussian Weight Distribution Model for Pixel Re-editing Details Alternate
  • Thomas - Moving Object Detection (OpenCV)
  • Kevin - Use webcam to take in video and separate into frames and output final image

References