Team Status Report
RISKS
Our greatest risk is the research aspect of our project. Riki will have a software implementation of seam carving shortly after spring break. Once that step is complete, we will be able to assess the efficacy of seam carving on different classes of videos. We also do not know exactly how long it will take to carve seams. That problem will be alleviated once the prototype is completed; however, until then, we have an unknown of how much time to allocate to that task in our design calculations.
DESIGN CHANGES
We have clarified the flow of data in our block diagram. For reference, the block diagram is copied below:
The preprocessing application will decompress the video into a 3 dimensional pixel array. It will be sent in this format to the FPGA over ethernet, where it will be written to RAM using DMA (one reason for the choice of ethernet is its DMA capabilities). The programmable fabric will read the video in this format, process the energy function, and write the resulting energies back to the RAM in the same 3D pixel array format. The SoC will read the energy values and compute the seam to remove. It will then perform a series of reads and writes to modify the video in RAM. The altered frames will need new energy values, so the programmable logic will recompute the energy function and update the energies stored in RAM. Once enough seams have been removed, the video will be transferred back to the computer in uncompressed format. The post-processing application will convert the video to a playable format and write the result to disk.
UPDATED SCHEDULE
Our schedule has been slightly tweaked to bring the prototype development earlier. The full Gantt chart can be seen here.
Maxwell Johnson
Personal work:
My main work this week was in organizing and writing the design report. I divided the design document into sections with desired dates of completion to organize the members of my team and to avoid leaving the document to the last day. The design report will be available on this site in a couple days, but I can post the division of labor as some evidence of this week’s work:
Earlier in the week, John and I worked on the slides for Monday’s presentation. We are revising them with feedback from the presentation in conjunction with our design report, but they will be available on this site within the week.
Progress:
Outside of class, my week was much busier than expected, so I did not complete the research on hardware that I was hoping to. My schedule is still aligned with our team’s Gantt chart, so I do not feel that my work on this project is behind.
Deliverables:
Before spring break, I will research the DDR3 controller and instantiate a test design on our FPGA. Depending on available time, I may also work with John on interfacing to the system on chip of our FPGA board.
John Zhang
Personal Work
This week, I and Maxwell prepared the slides for our team’s design presentation, rehearsed the presentation and I delivered the presentation. We got our feedback, and also discussed possible improvements with our assigned TA. I personally started going through periods of depression which is affecting my ability to contribute. To mitigate this, I have reached out to CaPs and scheduled appointments starting next week. I’ve also started my research on development on the SoC, as it is a key component of our system.
Progress
I’ve written some of the parts of the design report that I am responsible for. I found relevant courses and materials for SoC development and started reading into it.
Deliverables
For the next week I expect myself to improve our presentation slides by a lot, present it to Professor Low, and set up the environment for experimenting with the SoC remotely, since I would be away from campus for the duration of spring break.
Riki Khorana
Personal Work
This week, I mainly worked on setting up high-level requirements and their metrics. In particular, I came up with methods in which we would be able to assess our output video quality. Given my research, I revised my part of the gantt chart so I can work on the test suite effectively.
Progress
I made little to no progress on the actual implementation of my system. I created a rough prototype of decoding video data into bitstreams, to test the time it takes to complete the preprocessing.
Deliverables
By next Wednesday, I need to have the research of seam carving algorithms complete. This is so I can start implementing the software algorithm over the spring break.