Hi everyone, I've decided to detect brain cancer form an image library, so I wanna use the CNN algorithm to extract features from the library and train my neural network but still I wanna know how to choose the number of filters, filter shape, and max-pooling shape to have an optimized network? Thanks for sharing your answers.
One approach for hyperparameter optimization is grid search. There is a step by step tutorial for hyperparameter tuning:
https://machinelearningmastery.com/grid-search-hyperparameters-deep-learning-models-python-keras/
Meanwhile, as a separate thought, try finetuning a pre-trained Mask-RCNN (or YOLO). Along with classification, you will get localization too for brain cancer.
Dear Esmaeel Toni, Follow this CNN structure and you can set up parameters according to your involved dataset requirements (brain cancer, breast cancer, etc).
https://github.com/pytorch/vision/blob/master/torchvision/models/vgg.py
- Badges
- Science topic