This time, we are using PyTorch to train a custom Mask-RCNN. And we are using a different dataset which has mask images (.png files) as . So, we can practice our skills in dealing with different data types. Without any futher ado, let's get into it.

In this post, We will see how to fune-tune Mask-RCNN on a custom dataset. I will cover the processing pipeline from how to prepare a custom dataset to model funtuning and evaluation. It will be very useful, so keep reading.

Measuring foetal head circumference is one of the most accurate method to estimate gestational age after the first trimester (first three months of pregnancy). To automate this procedure, the most important step is segment the foetal head area in ultrasound images. Then we can fit an ellipse on the head area, and build a regressor using the head mask to estimate gestational age.
In this post, we will use a U-Net architecture with residual blocks to make head segmentations. U-Net performs very well on medical images. And due to its relatively compact structure, its speed is much faster than Mask-RCNN.

In this post, we are going to detect fraudulent transactions on credit card transaction data. This dataset can be downloaded here.

In this post, we will explore Mask-RCNN object detector with Pytorch. We will use the pretrained Mask-RCNN model with Resnet50 as the backbone.