KHAN MOHAMMAD AZAM is with the Data and Visual Analytics Lab in South Korea http://davian.korea.ac.kr/ and he will be travelling to the Endoscopic Artefact Detection Challenge 2019, hosted by Sharib Ali, thanks to a travel grant by MedIAN which has enabled him to travel and participate in this event, which is happening next week in Venice. He had been participating in the challenge prior to this but then, luckily for him, one travel grant recipient had to drop out so he got to go at no cost to himself because MedIAN likes to support early career researchers grow their network and career. He is a PhD student in the Department of Computer Science and Engineering (CSE), College of Informatics, Korea University, Seoul, South Korea having obtained his B.Sc.Engg. and M.Engg. in CSE from International Islamic University Chittagong (IIUC) and Bangladesh University of Engineering and Technology (BUET) respectively. His research interests include partial discharge and online condition monitoring of power system and big data analytics using deep learning techniques. We wanted to know more about how he felt about the challenge he'd just participated in and how he felt about the workshop in Venice that was coming up.

---

1) What do you think about EAD2019 challenge?
The EAD2019 challenge is a milestone for the growing application of artificial intelligence (AI) and deep learning (DL) techniques in the medical domain. Specifically, it is a unique challenge considering the diversity of the dataset for facilitating diagnosis and treatment of diseases in hollow organs. The organizers and the participants engaged very actively in this challenge for the last couple of months to tackle and solve some interesting problems in this field.

2) What was exciting about participating in the online challenge of EAD2019?
First of all, organizing an online challenge is not such an easy task. On top of that, collecting properly annotated medical data is another major issue nowadays. However, the organizers handled these challenges very nicely and accumulated the first large and comprehensive dataset, utilizing data obtained from six different data centers. It was really exciting to use such a wide dataset in the competition and get a chance to do some really diverse tasks as well.

3) what did you learn and what opportunities did it give you?
It was a nice experience for us to develop solutions for real-life problems. The competition provided us with an unprecedented platform in which we were able to compare our solutions, algorithms and model performance with each other. Such opportunities really help us to accelerate and understand the problem settings and probable solutions for this domain.

4) what excites you most about attending EAD2019 challenge?
It is still a challenging task to collect and label medical data reliably. It was really a great opportunity for the participants to work with a diverse dataset by taking part in this competition and I wanted to avail this opportunity by actively participating in the competition and in the subsequent challenge workshop. In addition, the tasks were also different from the usual machine learning competitions, I wanted to explore all the probabilities through this competition.

5) How do you feel regarding winning this challenge?
Honestly, I didn’t do very well in this competition. However, I think that many researchers and practitioners from the medical domain and healthcare sectors joined and competed in this challenge and that was a huge bonus for me. I feel really great for being one of the top ten successful participants in the challenge. In addition, MedIAN (Medical Image Analysis Network) is going to sponsor a travel grant which will help me to attend the post-challenge workshop smoothly in ISBI 2019. I am really thankful to MedIAN. 🙂

6) What do you think AI challenges like this can do to transform health care and help you understand the challenges and how can this be used for care pathways in the future?
Such challenges will encourage and enable close collaboration between AI/DL scientists and medical doctors. Then we may combine domain expertise with prediction/detection power of machines in an effective manner. Essentially, such challenges will help the participants to develop useful solutions which will transform healthcare by teaching computers to ‘see’ and ‘understand’ medical images through some interesting research and efforts.

It seems that the organizing committee was very engaging, cooperative and knowledgeable in this domain. Precise detection of specific artefacts and then precise boundary delineation of detected artefacts, and finally detection generalization of independent of specific data type and source – all would mark critical steps forward for this domain. Now it is the time to listen to the participants about their experiences, and at the same time, share my thoughts with others at ISBI 2019. I wish this challenge and the upcoming workshop every success.

Once again, thank you very much for organizing a great competition and selecting me for the travel grant. See you all at the EAD2019 workshop of ISBI 2019 in Venice, Italy.