TIA Centre: News

If a PhD is supposed to be a marathon, my second year felt more like a series of back-to-back sprints. As I finally catch my breath at the end of Year 2 here at the Tissue Image Analytics (TIA) Centre, I wanted to share a few stories from what turned out to be a wild, challenging, and incredibly fun ride.

Diving into the World of Grand Challenges

For anyone who hasn't had the pleasure (or madness) of participating in one, Grand Challenges are international competitions where labs like ours pit their algorithms against unseen datasets to solve a tough clinical or biological problem and being on an emotional rollercoaster. They are a true test of a model's robustness, generalisability, and whether your code actually works in the wild. This year, our team, TIAKong, decided to not just dip our toes in, but to cannonball into three of them.

Our adventure started with the MONKEY Challenge, which was about detecting and classifying mononuclear leukocytes in kidney biopsies. And here’s a funny thing: a project I did during my undergraduate dissertation supervised by Nasir at the time, which involved porting the NuClick model from TensorFlow to PyTorch, ended up being extremely useful here. At the time, to an undergraduate student it felt like an isolated piece of work. Who knew that this side-quest from the past would become a critical part of our data preprocessing pipeline and help us win a Grand Challenge? It’s a great reminder that you never know which skills or projects will come in handy years later.

Next up was the PUMA Challenge, tackling nuclei detection and tissue segmentation in advanced melanoma. This one forced me to think on my feet and get creative. I was able to integrate a massive pathology foundation model into a segmentation model to gain amazing performance. In both MONKEY and PUMA, TIAKong was absolutely buzzing to snag first place in one track and second in another.

The grand finale was the MIDOG 2025 Challenge, a huge benchmark for detecting and classifying mitotic figures, which have many useful applications in oncology. For Track 1, I was given the chance to lead team TIAKong once again, which was both an honour and terrifying. TIA had won the previous two MIDOG challenges, led by the legendary Mostafa Jahanifar, so... no pressure, right? I could definitely feel the weight of expectation. We were absolutely delighted to achieve first place in Track 1 (detection), and I was just as proud to see my colleague, Esha Sadia Nasir, brilliantly lead our team to a fantastic second-place finish in Track 2 (classification).

Those months were a blur of late-night coding and pouring over datasets. The intense brainstorming sessions weren't just about debating architectural choices; they were also about learning from others. For example, I could still remember the meeting with the amazing expert pathologist Brinder (Buzz) from University Hospitals of Derby and Burton NHS Foundation Trust, where he patiently tried to teach me how to identify lymphocytes in melanoma images. It’s moments like those, combining computational grit with clinical expertise, that make our work so powerful.

That spirit went beyond just one team. I also got to jump in on the CHIMERA Challenge, where our TIA-Pegasus team—led by the amazing Noorul Wahab and including Ethar Alzaid, Adam Shephard, and Shan Raza, stormed to first place in Task 1. There's something special about seeing different groups in the lab cheering each other on. It’s what makes TIA, TIA.

From the Lab to the World

This year wasn't just about competing from behind a screen; it was also about sharing our work and learning from researchers across the globe. I got to pack my bags and present my work at four incredible conferences, each in a city that seemed to tell its own story about science and innovation.

First up was ISBI in Houston. You can't be in Houston and not think about rockets, NASA, and the sheer ambition of space exploration. It felt fitting. The kind of "moonshot" thinking that sent people to space feels so connected to the ambition in our field.

Then came ECDP in Barcelona. While not a traditional tech hub, the city is a living museum of Antoni Gaudí's architectural genius. I was struck by his famous quote: "To do things right, first you need love, then technique." That perfectly sums up research. You need a genuine passion for the problem you’re solving, but that passion is powerless without the rigorous, deep technical skill to bring your ideas to life.

Closer to home, MIUA was in Leeds, a city that was a powerhouse during the Industrial Revolution. Being there was a cool reminder that we, in the AI community, are in the midst of our own revolution, one powered by data and algorithms instead of steam and steel.

The journey culminated with the big one: MICCAI in Daejeon. Often called the "Silicon Valley of South Korea," the city just buzzes with futuristic energy. You're surrounded by the sharpest minds pushing the absolute limits of what's possible.

From the moonshots of the past to the digital revolution of the future, each trip was more than just a presentation. Bringing the conversations, ideas, and inspiration from these places back to the lab felt like returning from an expedition with a hoard of treasure.

A Quick Thank You and What's Next

Of course, none of this happens in a vacuum. I owe a huge thank you to my supervisor, Shan E Ahmed Raza. He has this knack for encouraging us to take on ambitious projects while giving us the freedom to figure things out for ourselves. And a massive shout-out to the entire TIA Centre and the Department of Computer Science for providing the supercomputers that didn't crash (most of the time) and an environment where this kind of work can happen.

So, what’s the plan now? We’re not just letting the trophies gather dust. These challenge wins are the foundation for the final stretch of my PhD. The next chapter is about turning these competition models into solid scientific publications and, in the spirit of open science, getting our code out there for everyone to use.

This year taught me a ton about deep learning, sure, but it taught me even more about teamwork, leadership, and how to survive on coffee and sheer excitement. It’s been a blast, and I can't wait to see what Year 3 brings. Stay tuned!

By Jiaqi Lv