What do you do?
I implement and use AI algorithms to detect a wide variety of defects (>50) on lumber boards like knots, shakes, stains, wormholes, etc.
To be more specific, I use Convolutional Neural Network (CNN) to do so. CNNs are a class of deep learning neural networks. Deep learning neural networks are algorithms designed following the model of the human brain, and are used to recognize patterns. They are used effectively in image recognition and classification.
Image recognition is the ability to detect the object, classify, and recognize it. For example some
CNN are used to detect and recognize objects such as "car", "pedestrian", "truck", "motorcycle"
surrounding a self-driving car. The network first transforms the input image into a pixel-wise prediction image thanks to a semantic segmentation. The segmented regions are then classified into specific classes via a convolutional neural network.
Along with this, I also develop solutions for evaluating CNN models and then visualize the metrics generated from this evaluation.
The study of wood using machine vision is a challenge for many reasons, but all these reasons start with the fact that wood is a natural and biological material. Wood is not a homogeneous material. Each board is different. Between and within species, there may be hundreds of different characteristics and colors that are permitted as clear wood for grading purposes. In addition to the inherent variability of wood, there is also the coarseness of the rough lumber. Indeed, trees can be subjected to many different climatic conditions that lead to a wide range of defects.
Why did you choose this field?
I first wanted to become an Aerospace engineer so after high school I started 2 years of CGPE (French higher school preparatory classes that act as preparatory course with the main goal of training undergraduate students for enrollment in one of the Grande Ecoles.) where I learned a lot about Mathematics, Physics, Chemistry. I then realized I wanted to be able to work wherever I wanted and not only be tied to only one specific field. Thus, Applied Mathematics such as Computer Vision were perfect because it is used for aerospace engineering but also for medicine, cinema, and so much more! (At this time, I wanted to work for a medical imaging company.) So after these two intense years, I went to a school that has a Computer Vision major where I learned about image processing and machine learning. For my last Master semester, I did an exchange program at Oregon State University where I met the company I work for now that uses Machine Vision for wood analysis. I first thought it was weird and very Oregonian but now love doing this. Studying wood with machine vision is such a challenge, and even more than medical imaging sometimes.
What do you look at and think, "I wish younger me would have known this was possible"?
I wish younger me would have known that Computer Science is a tough field and can be very toxic especially when you are a woman but it does not mean that you can let people make you feel incompetent or not important, that working for a toxic and abusive company is not the only solution and that someday you will find a good company where people will listen to you and value your ideas. What is impossible though, is to work in a toxic environment and not stand up for yourself. It is hard, and it will take time, but someday you will be able to do it!
Why do you love working in STEM?
I love working in STEM because I can learn something new every single day. There are always challenges. I can spend hours and hours trying to debug a code or work with an AI model that does not want to learn what I need at all, and, yes, it drives me crazy but then when I find the solution, it's the best moment and I feel I have accomplished something new again.
Best advice for next generation?
STEM is fun! Computer Science is fun! We need more women in Computer Science! I see more and more women coming to this field and accomplishing so much because of they have a different vision of how things work. Also big plus, when you have Computer Science skills, you can work for basically any company and from everywhere in the world! You can be a part of projects linked to medicine, self-driving cars, wood science, climate change, business, website's designing, aerospace, aeronautics, biochemistry, biophysics, security, cinema, video games... anything you want actually!
Inspo quote / fun fact / role model
"Believe in yourself and go for it.”