We learned that detecting zero crossings by applying an operator that takes the second derivitive of an image is effective for edge detection (e.g., the Laplacian Operator). Yet we just learned that this approach isn't used in contemporary segmentation algorithms. Why not? First, explain why this is the case as the initial part of your answer. Support your answer by taking a photo of a problematic scene that you have identified / staged with your own camera and explain the aspects of the image that are confounding. To verify that that photo you've chosen as evidence is a good one, apply this OpenCV Laplacian operator code to the image and use the output as part of your answer. The Python environment you created for Practical 00 can be used for this.
To gain full credit for this homework, submit your written explanation, along with the original image you took plus the output of that image processed with the Laplacian operator, as a Google doc shared with Louisa.