Member-only story
Important CNN Concepts You Should Know in 2022
Hello, I decided to switch my focus a little bit and write something simple yet important for those who have just started their journey in Computer Vision. In this piece, I am introducing main and fundamental CNN concepts you need to know in 2022. I believe these concepts greatly contributed to Computer Vision community. Feel free to read more articles after this one since I tried to focus only on concepts rather than mathematical details. Thanks.
Normal Convolutional Layer
The image above shows well-know Convolution operation, I suppose we are familiar with. I do not want to focus on how it is computed since I believe you are already familiar with basic concepts if you’re reading this piece. The figure shows how kernels, gray cube, are applied on feature maps after which the resultant ‘yellow’ feature map is obtained. In this particular example, the input and output channels are 3 and 1, respectively, whereas the kernel size is 3. Also, I believe padding is applied since the size of the resultant feature maps is exactly same as the input. If some of the words sounds unfamiliar to you, please take your time to review basic concepts.
