i'm building a CNN to identify facial keypoints. i want to make the net more robust, so i thought about applying some zoom-out transforms because most pictures have about the same location of keypoints, so the net doesn't learn much.
my approach:
i want augmented images to keep the original image size so apply MaxPool2d and then random (not equal) padding until the original size is reached.
first question
is it going to work with simple average padding or zero padding? i'm sure it would be even better if i made the padding appear more like a background but is there a simple way to do that?
second question
the keypoints are the target vector, they come as a row vector of 30. i'm getting confused with the logic needed to transform them to the smaller space.
generally if an original point was at (x=5,y=7) it transforms to (x=2,y=3)- i'm not sure about it but so far manually checked and it's correct. but what to do if to keypoints are in the same new pixel? i can't feed the network with less target values.
that's it. would be happy to hear your thoughts
I suggest to use torchvision.transforms.RandomResizedCrop as a part of your Compose statement. which will give you random zooms AND resize the resulting the images to some standard size. This avoids issues in both your questions.
Goal:
For the past two weeks I've been trying to figure out how to convert the following image:
To one that looks like this (may not match exactly, as this image was taken at a different time):
Lens Correction (necessary?):
The first thing I noticed is that simply slicing the image and overlaying the four parts wouldn't work perfectly, as the curvature of certain lines does not match. For instance, the mid-court line bends left in the second slice and bends right in the third slice. This bending looks like a barrel distortion so I tried using both a parameterized lens correction function (passing k1, k2, and k3 to OpenCV) and using lensfun. Since the lensfun database does not include my camera make or model (it's an AXIS camera) and I do not know the make or model of the lens (it's manufactured as part of the camera), I wrote a small script to dump test images using various lenses with various parameters, then skimmed through the thousands of output images until I found one that looked like it had relatively straight lines:
This correction was done using the "Samyang 12mm f/2.8 Fish-Eye ED AS NCS" lens with a "Canon EOS 10D" camera in lensfun. It's probably not perfect, but I figured it was close enough to move on to step two.
Once the lens distortion was corrected, the second issue is that the same line in two slices was pointing in different directions, which should be corrected with a simple perspective transform. So I began a long quest to figure out the proper parameters for this perspective transform.
Failed Attempts:
1. Using SciPy
I started by writing a cost function to judge the "quality" of a given set of parameters (overlapped pixels should match) and applying SciPy's solver to figure it out. I made several tweaks to my cost function (applying a Gaussian blur, scaling down the image, gray scaling the image, using the Sobel operator to get a gradient, looking only at the pixels on either side of a "seam" after overlapping instead of the whole overlap region, etc) but it always failed to find a good solution. The results looked worse than the original camera image most of the time:
2. Using math
When that failed I tried applying math to compute the proper perspective transform. I know the FOV of the camera (from the spec sheet), I know the image width and height, I know the sensor size (from the spec sheet), and using a protractor I measured the angles between the lenses. Using the pinhole model I then calculated the expected (x,y) values of points on the image plane and what transform would be necessary to correct them. The results looked better than SciPy, but were still dismal.
3. Using OpenCV's Stitcher
After this I tried using OpenCV's built-in Stitcher class. However it failed to stitch together slices 2 and 3 due to insufficient overlap between the images (and about 10% of the time it even failed to stitch together slices 1 and 2, presumably because of the non-deterministic nature of RANSAC). Even when it did succeed, the stitch wasn't that great:
4. Using ORB and OpenCV's findHomography
Most recently I tried using ORB with a mask (only looking for features in the overlap region) and OpenCV's findHomography function to create a custom version of the Stitcher. While the matches seemed promising, the resulting stitch was still sub-optimal:
I'm beginning to suspect that my methodology (slice -> lens correct -> perspective transform -> overlay) is flawed and there's a better way to do this.
5. Updated ORB / findHomography
I updated my feature detection to eliminate any matches where the Y coordinates differed drastically (e.g. matching the white of the table to the white of the lights). After doing this my number of matched features fell from ~110 to ~55, but the homography was improved significantly. Here's the stitch that results for slices 1/2 and 2/3 with the update:
Until someone can tell me that I'm going about this all wrong, I'm going to keep pursuing this strategy with the following added step:
Slice image
Lens correct each slice
Perspective transform slice 2 or 3 so that the side line is horizontal and the mid-court line is vertical
Use ORB + match filtering + findHomography to iteratively align and then stitch adjacent slices
Ultimately when it's all said and done I want to try and compute a mapping from input pixels to output pixels so that we're not doing all of this complex work (lens correction, ORB, findHomography, etc) per-frame. We'll do it once per camera, save the mapping to a file somewhere, then we can in real-time map the input video to an output video frame-by-frame using cv2.remap
Note:
The second image I posted showing the "expected output" comes directly from the camera in question. It can be configured to return the first image at 30 fps, or the second image at 10 fps. We wish to perform the stitching off-camera on a more powerful computer so we can get 30 fps but still have the single image.
AXIS provides an SDK for doing the stitching off-camera, but this SDK is Windows-only and most of our tech stack is Linux and most of our development machines are Mac OS. I have used a Windows computer to try and look into the stitching SDK they provide, however I had no luck getting it to compile and run. Their sample code kept throwing errors and I've never had any luck getting Visual Studio or C++ to play nicely for me.
My suggestion is to train an autoencoder. Use the first image as input and the second one as an output, as in a denoising autoencoder:
Note that you may lose resolution if you create a botteleneck too small in the middle layer.
Also, Variational autoencoders present a latent vector but work following the same principle.
You can adapt this code:
denoise = Sequential()
denoise.add(Convolution2D(20, 3,3,
border_mode='valid',
input_shape=input_shape))
denoise.add(BatchNormalization(mode=2))
denoise.add(Activation('relu'))
denoise.add(UpSampling2D(size=(2, 2)))
denoise.add(Convolution2D(20, 3, 3,
init='glorot_uniform'))
denoise.add(BatchNormalization(mode=2))
denoise.add(Activation('relu'))
denoise.add(Convolution2D(20, 3, 3,init='glorot_uniform'))
denoise.add(BatchNormalization(mode=2))
denoise.add(Activation('relu'))
denoise.add(MaxPooling2D(pool_size=(3,3)))
denoise.add(Convolution2D(4, 3, 3,init='glorot_uniform'))
denoise.add(BatchNormalization(mode=2))
denoise.add(Activation('relu'))
denoise.add(Reshape((28,28,1)))
sgd = SGD(lr=learning_rate,momentum=momentum, decay=decay_rate, nesterov=False)
denoise.compile(loss='mean_squared_error', optimizer=sgd,metrics = ['accuracy'])
denoise.summary()
denoise.fit(x_train_noisy, x_train,
nb_epoch=50,
batch_size=30,verbose=1)
Introduction
I am newbie in statistics and NN.
I will describe my prolem below.
Problem
I want to find when the object on the video (images as well), will have following parameters:
the smallest diameter,
when the shape is closest to circle,
when the object have highest brightness.
Example:
https://imgur.com/a/pPIH6GX
Possible approaches
I started with neural networks (tensorflow) and object detection module, which detect objects similiar to circles, then I calculate the area of bounidng box. Then I started exploring mask RCNN networks and I am considering to obtain area of the mask. But it didn't solve my problem.
Asking for advice
The question is how to add other parameters? Like brightness or shape? Could you please give me some suggestions? Research papers, articles or post your thoughts in this field. I am waiting for deep conversation 😊
Many thanks!
Update 1
To detect circles Hough Transform can be used.
To find the smallest diameter I should consider thinness ratio.
Any other thoughts?
You should consider learning basics of image processing and computer vision first.
Locating circles on the image is a basic problem, so it's well-studied and has plenty of robust solutions.
There is no even need to use NN in your particular case (find a circle-like shapes at synthetic image you provided).
Refer to Wikipedia enter link description here to start with basics.
I am trying to obtain a radius and diameter distribution from some AFM (Atomic force microscopy) measurements. So far I am trying out Gwyddion, ImageJ and different workflows in Matlab.
At the moment the best results I have found is to use Gwyddion and to take the Phase image, high pass filter it and then try an edge detection with 'Laplacian of Gaussian'. The result is shown in figure 3. However this image is still too noisy and doesnt really capture the edges of all the particles. (some are merged together others do not have a clear perimeter).
In the end I need an image which segments each of the spherical particles which I can use for blob detection/analysis to obtain size/radius information.
Can anyone recommend a different method?
[
I would definitely try a Granulometry, it was designed for something really similar. There is a good explanation of granulometry here starting page 158.
The granulometry will perform consecutive / increasing openings that will erase the different patterns according to their dimensions. The bigger the pattern, the latter it will be erased. It will give you a curve that represent the pattern dimension distributions in your image, so exactly what you want.
However, it will not give you any information about the position inside the image. If you want to have a rough modeling of the blobs present in your image, you can take a look to the Ultimate Opening.
Maybe you can use Avizo, it's a powerful software for dealing with image issues, especially for three D data (CT)
I would like to find a zoomed microscopy image in a dozen of overview images. I would prefer to find some python/numpy/scipy solution.
My knowledge about pattern recognition is negligible. Anyway, here is what I tried:
My first idea was to get the most important structures out of the images by
setting eveything greater[smaller] than some threshold to 255[0] in the greyscaled image.
For example, I then have the following pattern:
The overview image might then look like this:
Here is a version where the region of the pattern is highlighted:
I would now like to find a way to get the pixel number, at which the pattern occurs in the overview image.
It is very important to note, that I do not have information about:
the orientation of the two images with respect to each other
the scaling of the images
in principle, there might even be some stretching between the images, but this might be too hard to implement.
I do no know if the pattern is on the image. There I have to check about 20 images.
For a fixed scaling, my attempt was to use
result = scipy.signal.fftconvolve()
and see how the maximum of result then varies when I rotate the pattern before doing the convolution.
Taking the maximum value gives me the correct angle at which the images overlap.
However this is not a nice solution because it already takes some minutes. Also varying the scaling and even doing further transformations would take forever.
I guess there are better approaches out there!