Systems and methods for training a machine-learning algorithm and application of a machine-learning model for denoising images
A system for training a machine-learning algorithm for denoising images is configured to receive training data. The training data includes multiple image sets obtained from one or more series of consecutive images. Each image set includes a plurality of images obtained from a same series of consecutive images. The plurality of images of each image set includes an initial image, a middle image and a last image. The system is further configured to adjust weights of the machine-learning algorithm to obtain a trained machine-learning model, based on an output image of the machine-learning algorithm and a target image. The output image is obtained by using the initial image and the last image as input images. The target image is obtained from the middle image by applying a random shift to the middle image. The system is further configured to provide the trained machine-learning model.
1 . A system comprising one or more processors and one or more storage devices for training a machine-learning algorithm for denoising images, the system configured to:
receive training data, the training data comprising multiple image sets obtained from one or more series of consecutive images, wherein each image set comprises a plurality of images obtained from a same series of the one or more series of consecutive images, wherein the plurality of images of each image set comprise an initial image, a middle image and a last image;
adjust weights of the machine-learning algorithm to obtain a trained machine-learning model, based on an output image of the machine-learning algorithm and a target image, wherein the output image of the machine-learning algorithm is obtained by using the initial image and the last image as input images, and wherein the target image is obtained from the middle image by applying a random shift to the middle image; and
provide the trained machine-learning model.
2 . The system of claim 1 , wherein adjusting the weights of the machine-learning algorithm to obtain the trained machine-learning model, based on the output image of the machine-learning algorithm and the target image is performed by minimizing a difference between the output image and the target image.
3 . The system of claim 1 , wherein the plurality of images of each image set are consecutive images of the respective series of consecutive images.
4 . The system of claim 1 , wherein the one or the more series of consecutive images are obtained from at least one of: time lapse imaging, or one or more videos.
5 . The system of claim 1 , wherein the training data comprises noisy images.
6 . The system of claim 1 , wherein the machine-learning algorithm comprises a convolutional network or a deep convolutional network.
7 . A computer-implemented method for training a machine-learning algorithm for denoising images, the method comprising:
receiving training data, the training data comprising multiple image sets obtained from one or more series of consecutive images, wherein each image set comprises a plurality of images obtained from a same series of the one or more series of consecutive images, wherein the plurality of images of each image set comprise an initial image, a middle image and a last image;
adjusting weights of the machine-learning algorithm to obtain a trained machine-learning model, based on an output image of the machine-learning algorithm and a target image, wherein the output image of the of the machine-learning algorithm is obtained by using the initial image and the last image as input images, and wherein the target image is obtained from the middle image by applying a random shift to the middle image; and
providing the trained machine learning model.
8 . A non-transitory computer-readable medium storing a trained machine-learning model for denoising images, wherein the machine-learning model is trained using one or more processors by executing the follow steps:
receiving training data, the training data comprising multiple image sets obtained from one or more series of consecutive images, wherein each image set comprises a plurality of images obtained from a same series of the one or more series of consecutive images, wherein the plurality of images of each image set comprise an initial image, a middle image and a last image; and adjusting weights of a machine learning algorithm to obtain the trained machine-learning model, based on an output image of the machine-learning algorithm and a target image, wherein the output image of the of the machine-learning algorithm is obtained by using the initial image and the last image as input images, and the target image is obtained from the middle image by applying a random shift to the middle image.
9 . A system comprising one or more processors and one or more storage devices, for generating a denoised image by applying a machine-learning model, wherein the system is configured to:
receive input data, the input data comprising two input images obtained from a series of consecutive images,
apply the machine-learning model to the two input images in order to obtain a denoised image; and
provide output data, the output data comprising the denoised image,
wherein the machine-learning model is trained by:
receiving training data, the training data comprising multiple image sets obtained from one or more series of consecutive images, wherein each image set comprises a plurality of images obtained from a same series of the one or more series of consecutive images, wherein the plurality of images of each image set comprise an initial image, a middle image and a last image; and
adjusting weights of a machine learning algorithm to obtain the trained machine-learning model, based on an output image of the machine-learning algorithm and a target image, wherein the output image of the machine-learning algorithm is obtained by using the initial image and the last image as input images, and the target image is obtained from the middle image by applying a random shift to the middle image.
10 . The system of claim 9 , wherein the two input images are two copies of an individual image.
11 . The system of claim 9 , wherein the two input images are two different images, and wherein the denoised image is an interpolated image based on the two input images.
12 . The system of claim 10 , wherein the two input images are two consecutive images of the series of consecutive images.
13 . The system of claim 9 , wherein the two input images are obtained from time lapse imaging or from a video.
14 . The system of claim 9 , wherein the two input images are obtained from time lapse imaging of one or more living samples.
15 . A computer-implemented method for denoising an image by applying a machine-learning model, the method comprising:
receiving input data, the input data comprising two input images obtained from a series of consecutive images,
applying the machine-learning model to the two input images in order to obtain a denoised image; and
providing output data, the output data comprising the denoised image,
wherein the machine-learning model is trained by:
receiving training data, the training data comprising multiple image sets obtained from one or more series of consecutive images, wherein each image set comprises a plurality of images obtained from a same series of the one or more series of consecutive images, wherein the plurality of images of each image set comprise an initial image, a middle image and a last image; and
adjusting weights of a machine learning algorithm to obtain the trained machine-learning model, based on an output image of the machine-learning algorithm and a target image, wherein the output image of the machine-learning algorithm is obtained by using the initial image and the last image as input images, and the target image is obtained from the middle image by applying a random shift to the middle image.
16 . The method of claim 15 , wherein the two input images are two copies of a single image.
17 . A non-transitory computer-readable medium having a program code stored thereon, the program code, when executed by a computer processor, causing performance of the method of claim 7 .