Some suggestions for training a segmentation network are using “unet” and “deeplabv3plus” models.
DeepLab v3+ uses atrous (or dilated) convolutions, which allow the network to capture multi-scale contextual information without losing resolution. This is particularly useful for segmenting objects at various scales.
Please refer to the below link for training a segmentation model using “unet” model.
For more information on how to train a segmentation model using “deeplabv3plus” refer to the following link.
Please note that you need to have a MATLAB R2024a or later version to implement the above-mentioned segmentation models.
By using the above segmentation models, there will not be any need to extract the features of image separately.
Hope you find this information helpful.
