Community Profile

photo

Sourav Bairagya


Last seen: Today

MathWorks

62 total contributions since 2019

I am an Engineer in the Engineering Development Group here at MathWorks.

My interest lies in the field of Image Processing, Computer Vision and Deep Learning.

DISCLAIMER: Any advice or options here are my own, and in no way reflect that of MathWorks.

Sourav Bairagya's Badges

  • 3 Month Streak
  • Revival Level 2
  • Knowledgeable Level 3
  • First Answer

View details...

Contributions in
View by

Answered
How to add batch normalization layers in between Con. layer and RELU layer in Googlenet.? Suggestions to improve accuracy..
To add new layers in layergraph object, first add the new layer using 'addLayer' function. Then, you can use 'connectLayers' fun...

10 months ago | 0

Answered
How to perform Faster RCNN on my own dataset in Matlab?
After you prepared ground truth from your dataset, load that ground truth data. Now, extract the training data from that ground ...

10 months ago | 0

Answered
Calculating the sum of each row in a matrix.
You can use 'sum' function to compute sum for each row and then can apply 'squueze' function to remove the singleton dimension (...

10 months ago | 0

Answered
trainNetwork error RESHAPE line 260
It seems that the error lies in using 'reshape' function in 'Crashworthiness_NN'. It may be possible that the number of elements...

10 months ago | 0

Answered
Using a trained Faster RCNN Object Detector, expected LABEL to be nonempty
It seems that the labelling is not done properly. It may be possible that some images are not labelled. So, it would be good t...

10 months ago | 0

Answered
how can one utilize a dropout layer in a neural network during prediction?
Usually dropout layers are used during training to avoid overfitting of the neural network. Currenly, 'dropoutLayer' of 'Deep le...

10 months ago | 0

Answered
Monitor Deep Learning Training Progress
You can leverage these links to get idea about how to train multiple deep learning networks and observe their training progresse...

11 months ago | 0

Answered
In the GAN example of the documentation, did we update D several times before updating G?
GANs can be trained in many fashions. Here, discriminator, D is updated once before updating generator, G. Hence, first, discrim...

11 months ago | 0

Answered
how can I do object detection with vgg16 by using rcnn??
You can extract the desired number of layers from the pretrained net and define your choice of network. Here is an small example...

11 months ago | 0

Answered
Create geometry from mesh surface mesh
You can leverage these links: https://www.mathworks.com/help/pde/ug/pde.pdemodel.geometryfrommesh.htmlsearchHighlight=geometryF...

11 months ago | 0

Answered
ADAS// Lane change
As the error suggests that it is possible that the block supports variable size signals but needs to be configured for them, you...

11 months ago | 0

Answered
File Doesn't contain valid data of type 'SignalLabelDefintion'
You can leverage this link to get more details on the import and export behaviour of the signal labeler: https://www.mathworks....

11 months ago | 0

| accepted

Answered
Forward outputs from discriminator of GAN vary with the number of inputs.
It seems that the discriminator layer weights get chnaged during second time calling of "forward" function. Make sure that the d...

11 months ago | 1

| accepted

Answered
How to make a simple gui for image processing?
You can leverage these links for creating apps using App Designer: https://www.mathworks.com/help/matlab/app-designer.html htt...

11 months ago | 0

Answered
What is the mean of "Error using builtin" when using trainNetwork?
It seems that data stored in the '.mat' file may have some error. Hence, it would be better to check the dimesnions of those dat...

12 months ago | 0

Answered
Accessing dlarray elements for further processing
You can follow this approach, A = randn(6,6); A = dlarray(A); A1 = A(1:3,:); A2 = A(4:6,:); Here, A is a matrix of 6X6 dime...

12 months ago | 0

Answered
trainning multiclass in rcnnObjectDetector
You can use 'Image Labeler' app to label your training data. For more information, you can leverage this links: https://www.mat...

12 months ago | 0

Answered
After creating YOLO v2 with resnet50, while training the network, getting error "Invalid training data. The output size ([7 7 16]) of the last layer does not match the response size ([1 1 3])."
From the error it seems that output of your network should have size [1 1 3], but it is having the size [7 7 16]. May be there i...

12 months ago | 0

Answered
How can I make a meshgrid of this image? I am trying to create a bathymetry with matrix.
You can do that following this way. Assume the image is I. [X,Y]=meshgrid(1:size(I,2), 1:size(I,1)); mesh(X,Y,I) %display mes...

12 months ago | 0

Answered
down sample nearest method
The command 'imresize(im,0.5,'nearest')' will first find out the query points i.e. it will map the output image cordinates to th...

12 months ago | 0

Answered
supervised learning to transform multi-dimensional data to 2D?
You can build your custom neural network to learn this mapping function from 14 dimensional data to 2 dimensional data. In this...

12 months ago | 0

Answered
How can I add gaussian noise to image that are of class double in the range [0, 255]?
The 'imnoise' function expects input pixel values of data type 'double' and 'single' to be in the range [0, 1]. It also accepts ...

12 months ago | 0

| accepted

Answered
Shoud I preprocess the input data before using fitcsvm?
It is better to preprocess the input data before using 'fitcsvm'. There is option to standardise your input data by setting the...

12 months ago | 0

| accepted

Answered
cnn for feature extraction
As you have your RGB images ready, then you can define your custom convolutional neural network using 'dlNetwork' object and tra...

12 months ago | 0

Answered
Regarding 3D-UNET for segmentation.
Dataset for training 3D U-Net contains 4-D voulmes of which each one represents a stack of 3D images. The first three dimensions...

12 months ago | 0

| accepted

Answered
Regarding GAN and its loss objective
You can use 'mse' function from Deep Learning Toolbox which calculates half mean sqaured error between given two inputs. dlY = ...

12 months ago | 0

| accepted

Answered
how to change the color output of segmentation?
For "imshowpair" function, inputs should be grayscale/ binary/ truecolor images and for "jaccard" function, inputs sholud be bin...

12 months ago | 0

Answered
Error using Semantic Segmentation Using Deep Learning
I have tried to replicate that same example (mentioned in the provided link) on MATLAB R2019b and it sucessfully ran. On MATLAB...

12 months ago | 0

Answered
image labeler_Label Pixels Image
You can use "object DetectorTrainingData" function to create table form the labelled ground truth objects which are returned fro...

12 months ago | 0

Answered
How to TRAIN further a previously trained agent?
In this case, you can resume your training with the previous experience buffer as a starting point. You have to set the 'SaveE...

12 months ago | 0

Load more