Starting his career in computer vision after doing his thesis at SICK

We are so proud of the students who come do their thesis at SICK, and the journey they have afterwards. Abhijeet Anand graduated last summer, so we asked him to talk about his experience: what made him choose to do his thesis at SICK Linköping, what it was like, and what happened after graduation. An inspiring journey of diving into machine vision and deep learning. 

A student in statistics and machine learning, Abhijeet Anand wanted to focus his career on machine vision. While his program did not offer related courses, he decided to do his thesis at SICK on this topic. The decision was a risk, but it ultimately paid off: by the time he graduated, he was offered a role as a computer vision engineer in a Malmö-based startup.

Can you introduce yourself?

I am from India, where I did my bachelor's in Electronics and Communication back in 2018. I then joined Siemens as a software developer. During a hackathon organized by Siemens, I in a team of 4, developed something related to AI. It was helpful for Siemens to discover more about the topic, and it generated some initial interest for me towards AI and machine learning. That was when I started looking for a Masters - ultimately leading me to my master’s degree in statistics and machine learning at Linköping university.

How did you decide to do your thesis at SICK?

I attended the thesis night at SICK. Before that I didn’t have a lot of knowledge about SICK IVP, but I saw the event online and registered for it. At the event, I met the team who introduced us to all the projects. I applied for three projects that were the most interesting to me.

In the end, I had four thesis offers including the one that SICK offered me, but I ended up choosing SICK. First, because the idea of the thesis was very appealing as I wanted to focus on deep learning, AI, and machine vision. Second, because of my supervisor, Christopher. During my interview at SICK, I really liked his thought process and approach towards the problem statement. I felt he would be a really nice supervisor for my thesis.

I chose to do my thesis at SICK, first because the idea of the thesis was very appealing as I wanted to focus on deep learning, AI, and machine vision. Second, because of my supervisor (...) I really liked his thought process and approach towards the problem statement.

What was your thesis about?

My thesis title was “Point cloud registration in the application of bin picking”.

The problem statement says - there would be around twenty known objects in a small box, jumbled upon each other. A camera placed on a robotic arm captures a 2.5D image of it - so a picture with depth sense. Point clouds of the same image are generated. The task was then to detect individual objects from these piled up objects, followed by calculating the 6D pose of the detected items.

It is challenging in the sense that all the objects are the same: if you have different shapes, it becomes easier to segment it, but when they are identical, it is difficult. So, the first part of the task was to detect individual objects - technically, to perform 3D instance segmentations. Then calculate its angle in 3D space by performing Point cloud registration. This research was carried out to improve the grasping strategy of a robotic arm. If the 6D pose of the object is known, the robotic arm can pick up the object with a better grip.

How did you approach it?

The topic was very challenging, and I had no computer vision experience. Moreover, in my master program we have a hard deadline - it is not possible to extend the thesis deadline at all. Which is why I started working on it a month in advance, in agreement with SICK.

Even so, for the first two weeks I wondered why I chose this topic rather than an easier one! But in the end, it paid off. The best part of my thesis though was my supervisor. We did not discuss a lot, but consistently - about every day about 10-15 minutes. Those inputs were very helpful. He also had an open-door policy: whenever I had an issue, I could come to him to talk. He was always available for me, and it was a very good experience.

"My supervisor had an open-door policy: whenever I had an issue, I could come to him to talk. He was always available for me, and it was a very good experience" 

Did you speak Swedish when you started? What was it like for you as an international student?Nej, jag pratar inte svenska - eller bara lite! I don’t speak Swedish, but the environment at SICK did not force me to speak Swedish. I know it can be a question for international students, but for the entire time of my six months there, everyone was friendly and easy to communicate with in English.

What happened after graduating? How was the transition to professional life?

I really wanted to center my career towards AI and deep learning with a focus towards computer vision. Towards the end of the thesis, I got interviewed for a company in Malmö, Blue Cromos. They hired me for a computer vision role.

Since I did not have any course in computer vision during my master’s degree, the only experience I had with computer vision is during the thesis at SICK, so I give credits to my time here in getting this position.

Can you talk more about your current work? What are the best parts, and your current challenges?

It’s a startup company working with anti-counterfeiting and other computer vision solutions. They work with protecting brands from fake products. The challenge of fake products can have quite high stakes: for instance, if you get any fake parts in an airplane, it can put many lives at stake.

The best part is that it is a startup: when the company is small, you get to know everyone. You don’t have any communication gap and it is easy to talk to anyone. I also enjoy the work very much. The problem itself is challenging, so I have to do good research, which I am happy about. There is nothing repetitive, and I get to work with many different things at the same time.

The only challenge is that Malmö is a new city for me: rebuilding your social network is a challenge everytime you move. Also, I miss Linköping, and I am especially going to miss the snow in winter!

What is your best memory from the time of doing your thesis at SICK?

Should I talk about work or fun?

Both!

On the fun side, there was a table tennis tournament during my time here - and I really love to play table tennis. I went to the semi-finals, but I couldn’t make it to finals. But there was one fun side event to it: the fastest smash challenge. I am sure people at SICK will still remember me for that smash! I jumped, smashed… and my bat broke! It flew right in front of the CEO, leaving me with only the handle in my hand. I did win the fastest smash challenge with that hit though!

For a more thesis-related memory, one thing which I liked is that I was sitting in the thesis area in the deep learning team. In that space, there is a big white board. As I am very organized, I like to plan and write a lot. Since no one was sitting beside me, I had this entire board just for me: I ended up painting it in every corner! It helped me plan and organize my thesis better.

What advice would you give to other students considering doing their thesis at SICK?

Doing a thesis at SICK would be a very good decision because the company in general is very helpful, and everyone is very communicative. You can approach anyone with your doubts. It is a known fact that when you start the thesis it is very challenging, but as you keep going, things start getting into shape. The requirements are very high, and people want you to go the extra mile, but SICK offers very interesting topics, and the people are very nice and helpful.

"The company in general is very helpful, and everyone is very communicative (...) It is a known fact that when you start the thesis, it is very challenging, but as you keep going, things start getting into shape. "