Questions tagged [mask-rcnn]
For questions about the mask R-CNN model, proposed in the paper "Mask R-CNN" (2018) by Kaiming He et al.
14
questions
0
votes
0
answers
12
views
Is it a good idea to have a category and its subcategories in the training set of an object segmentation model?
I am currently training an object segmentation model (detectron2 : mask rcnn)
The objective is to detect materials like wood, plastic, glass etc...
wood is one of the categories in my training set.
Is ...
0
votes
0
answers
27
views
Should the number of training iterations of an instance segmentation model depend on the number of instances in the training dataset?
I need to train instance segmentation models on several different datasets. The datasets vary widely in how many instances each image contains.
For example:
...
- 1
1
vote
1
answer
31
views
possible to combine multiple labeled objects as one object?
So I have labeled the entire skeletal muscles in detail. For example instead of just labeling shoulders I have labeled:
Rear Delt
Middle Delt
Front Delt
but now you want all of the delts to be ...
- 11
0
votes
1
answer
24
views
Why instance segmentation architectures using reconstruction masks but not regression?
I'm wondering why many model architectures use binary mask reconstruction for segmentational CNNs, and not regression of mask polygon coordinates? Many object detectors use regression to find ...
- 103
0
votes
0
answers
19
views
is there any variation in the results if you resize a image with black lines?
Hello I need to resize a lot of images each of these has its own random size, for example, I have photos with the following size
100x200
102x200
202x201
...
in general, the resolution of each photo of ...
- 113
2
votes
2
answers
104
views
How to instruct Mask RCNN to identify objects too close to each other?
I have been trying to train a Mask RCNN model to identify individual poker chips in a stack. No matter what property I change, the end results look like the following image. I was guessing the issue ...
- 29
1
vote
1
answer
300
views
Mask R-CNN: How are the computed masks projected back to the input image?
The computed masks by Mask R-CNN are of fixed size $m \times m$ each. How are they projected back to the image?
- 21
1
vote
0
answers
242
views
Mask R-CNN: how is the inference done?
According to the Mask R-CNN paper and the picture below (taken from the paper), the mask branch is computed in parallel with the bbox classification and regression branches.
However in the paper they ...
- 21
1
vote
0
answers
31
views
Can I use a Mask R-CNN to detect a skin texture?
I'm trying to implement a solution in python to detect skin in an image.
I'm evaluating the Mask R-CNN model to create a mask on the skin (not on clothes). The problem is that every solution I have ...
- 11
4
votes
1
answer
997
views
How does Mask R-CNN automatically output a different number of objects on the image?
Recently, I was reading Pytorch's official tutorial about Mask R-CNN.
When I run the code on colab, it turned out that it automatically outputs a different number of channels during prediction. If the ...
- 852
0
votes
0
answers
96
views
Improving Mask RCNN by arbitrary scaling head input
Currently, I am looking at how Mask R-CNN works. It has a backbone, RPN, heads, etc. The backbone is used for creating the feature maps, which are then passed to the RPN to create proposals. Those ...
- 101
2
votes
0
answers
532
views
Getting bounding box/boundaries from segmentations in UNet Nuclei Segmentation
From my understanding, in a tissue where nuclei are present and need to be detected, we need to predict bounding boxes (either rectangular/circular or in the shape of the nucleus, i.e. as in instance ...
- 159
2
votes
0
answers
1k
views
How much should we augment our training data?
I am wondering how much I should extend my training set with data augmentation. Is there somewhere a pre-defined number I can go with?
Suppose I have 10000 images, can I go as far as 10x or 20x times, ...
- 21
2
votes
1
answer
831
views
Is Mask R-CNN suited to solve a multi-class classification problem where the classes are related?
I want to create a model to solve a multi-class classification problem.
Here are more details about my problem.
Every picture contains only one object
The background is very simple
All objects ...
- 21