Introduction
On 13th December 2019, I have joined Machine Learning with Python mainly for object detection conducted by Miss Dalila Al-Husna and her team from Synapse Innovation Sdn Bhd (sub-company of UTM). The workshop was held at Lab Teaching 1 N28A.
Program Outline
- The CompFair's team help the participant to install the necessary program in the laptop which is python 3.8, cascade-trainer GUI, and notepad++
- Then they gave us material for our cascade to train which consists of 500+ negative and positive images. The negative images are images of the road while positive images are images of the pen.
- The instructor reminded us to use the cascade trainer we need to have a set of more than 50% of positive images. For example, if we want to use 300 positive images, we need at least 600 negative images to train the datasets.
- After training the dataset, the cascade trainer output a .xml file. The file then is put inside a python file for image detection using a laptop's camera.
- It then can provide a live feed of the program detecting an object in this case, pen.
Detected as PEN!
.xml file after training
After the program
Conclusion
The CompFair's team was very supportive and passionate about their work. I had learned a lot from this program. The machine learning is one of the method use for Artificial Intelligence and proven to bring goodness to mankind. This workshop has inspired me to learn more about AI and make me confident in my future choice to proceed in an AI path.