Community Bonding Week-1

I am excited to be part of GSoC 2025! This week marks the official start of the program with community bonding period, and I couldn’t be more thrilled to dive in.

My project, “Extend DetectionMetrics: GUI, CI Workflow, and Object Detection”, will introduce me to perception model evaluation and give me hands on experience in the field of computer vision. DetectionMetrics is a toolkit used for evaluating perception models across various frameworks and datasets. Ironically the latest version of Detectionmetrics currently supports image segmentation model evaluations, during my coding period my objective is to extend its functionality to object detection evaluation, build an interactive user interface and set up the CI workflow. This project is a perfect opportunity for me to gain hands-on experience in computer vision, and I’m eager to soak in all the learning that comes with it!

---

Meeting with Mentors

This week, I had my first meeting with my mentors. We spent time getting to know each other and discussing the initial steps of the project. We also reviewed the program’s logistics and clarified expectations moving forward. Communication was emphasized, with Slack chosen as the primary tool for day-to-day discussions and updates.

The discussion also covered the initial steps for the project, such as diving deep and learning about different perception models, standard frameworks and datasets. I was also asked to set up this blog to document my journey through the project and for it to be easy for my mentors to keep a track of my progress.

This Week’s To-Do

• Set up a blog website using Jekyll and GitHub Pages
• Read and analyze literature on perception models and datasets
• Write this first blog post documenting progress

---

Progress So Far

I started the week by solidifying my foundation, I revisited my undergraduate computer vision coursework, focusing on:

• Key concepts: Object Detection, Instance Segmentation, and Semantic Segmentation
• Core frameworks: PyTorch and TensorFlow
• Standard datasets: COCO, Pascal VOC
• Evaluation metrics: mAP, IoU

This foundational review has been super helpful in framing the work ahead and getting me ready for the development phase.

Then I moved on to setting up this blog — drawing inspiration from past GSoC contributor blogs, I used Jekyll , a static site generator, along with GitHub Pages for hosting the website. I plan on uploading a blog every week with the progress, updates and learnings.

---

I’m looking forward to continuing this journey, learning from my mentors, and contributing meaningfully to the open-source community.