Open CVC++ Language
1- For the first section, you are required to recognize the droplets from a video using computer vision algorithms that is taken by a high-speed camera during the droplet 3D printing process. The video can be found from the attached below:
Find below for the detailed tasks of the topic you need to follow, indicative effort in brackets:
• You will need to recognize the inner droplet.
• You will need to recognize the outer wrap.
• You will need to count for all the successfully formed droplets.
• You will need to do centre of mass/blob detection.
2- Following above, it will be for NN tasks, r2222222222222222t022222220 0 0 shown below, trained by the CIFAR-10 dataset, to classify the images given by me.
Write a technical report for this with the (a) demonstration of its theorical contents, maths, etc;
(b) by adjusting the parameters, analysis in what situation, the system has best recognition performance;
(c) compare your designing with the one in Practical 9, write down the critical analysis for it.
(d) It is designed with MLP convolutional layers and propose global mean pooling, you need to assure the accuracy is over 90%. Its architecture is given in a table below, which you need to refer for the design of your NN model:
1- Writing: description of the methods you used, theory on code implementation, use of English and Figures
2- Mathematics: math equations will be your model of the system, maths in your coding & report need to be both used and presented correctly
3- Source Code: quality, extent and commentary.