WEEK 8: TRY DIFFERENT OBJECT DETECTION METHOD & LEARN DEEP LEARNING FOR COMPUTER VISION

Date: 1 - 5 May 2023

Content:
In week 8, I was focused on expanding my knowledge of deep learning for computer vision and finding a suitable alternative for object detection in my Final Year Project (FYP). I learned machine learning basics, classification metrics, neural networks, gradient descent, Keras, MNIST dataset and convolutional neural networks. 

In this week 8, I encountered some challenges with the 'Detecto' library, as its detection speed did not meet the real-time requirements of my project. As a result, I explored new method on object detection called 'Yolov5' known for its speed to detecting object and I utilized 'Roboflow' annotation tool to support the training process.

Conclusion:
Overall, week 8 involved a deep dive into deep learning for computer vision. Furthermore, I made an important decision to shift from 'Detecto' to 'Yolov5' object detection method due to its improved speed.

Appendix:
Figure below shows the output for fish detection by using 'Detecto'

Figure: Detecto Output


Comments

Popular posts from this blog

WEEK 3: IMAGE BASICS AND OPENCV

WEEK 6: LEARNING OBJECT DETECTION FOR MY SYSTEM

WEEK 2: FYP WORKSHOP & COMPUTER VISION INTRODUCTION