Graduate STEM Fellow Profile
IMPACT LA: Improving Minority Partnerships and Access through CISE-related Teaching
Thesis: Motion Estimation Using FPGA-based Architecture with Specialized Processing Elements
College/University: California State University, Los Angeles
Research Advisor: Dr. Charles Liu
Degree Sought: Masters in Electrical Engineering
Department: Electrical/Computer Engineering Lab
Research Focus: Efficient ways to map data to calculation units on FPGA (field programmable gated array). Apply motion estimation algorithm to t
Teaching Partner(s): Israel Hernandez, Adriana Trejo Sanchez
Description of Research
Video compression can be achieved using a key technique called motion estimation. In motion estimation, we can exploit the spatial and temporal redundancies by comparing the previous video frame with current video frame to determine if any of the objects between two frames has any movement between them. A motion vector that specifies the displacement of objects between current frame and previous frames along with pixel information can be sent from the system rather than video frames with complete video information which contains more data bits. Motion estimation is the most computational intensive part of the video compression. It alone can take up to 65% of total computational resources. High memory utilization due to memory overlapping and data-stream reuse is another common problem. By using FPGA (field programmable gated array) with specialized processing element hardware, parallel processing, data-stream mapping/special memory organization and bit-reduction along with object boundary detection to reduce computational complexity can help to achieve desirable calculation performance and video visual quality.
Example of how my research is integrated into my GK-12 experience
The iteration design process used in engineering was introduced into the classroom where an activity that encourages the students to come up with unique design by using simple everyday household products. The initial ideas are very rough and not very refined. The iteration process assured the students that doing things repeatedly until it is in an acceptable state is not waste of time really paid off.
Other examples where my research is integrated into the classroom is that an introduction presentation about how computer represent colors and images and how computer represent data that got the students interested in the primary colors and binary numbers. The students learned about how pixels and colors are represented in the computer and we related these concepts to the resolutions of the digital camera.