Chi Nguyen’s status report for 02/18/23

As mentioned in my previous status report and in our schedule, I have been working on implementing YOLO algorithm for object detection process. I have created a github which includes all of my code and notes on the design and implementation process (you can find it here: https://github.com/cnguyen09/NoTimeToDine). I spent lots of time working on understanding different steps in the algorithm and its design portion so I’m expecting to push more code to our git repo later tomorrow after our group’s meeting. I am currently on schedule as I’m expecting to finish this piece of code by the middle of next week or the end of next week at the latest. Currently, I’m aiming to have my code able to work on a simple set of images that include one object each. Once the algorithm works well on the simple dataset, I’ll add more layers in my code using Object-oriented programming so that the code works on images with multiple objects.

Regarding the engineering science and mathematics principles in my design, I mentioned that I would use a statistical modeling for the wait time estimation portion of the project which I’ve learned in 36-226: Introduction to Statistical Inference. In terms of object detection, YOLOv7 uses a re-parameterization convolution method called RepConv. I am very familiar with different convolution methods as I learned them from 18290 and 16385 Computer Vision so I find myself picking up RepConv pretty quickly. Here is a published paper on everything that I need to know: https://arxiv.org/pdf/2207.02696.pdf