An official website of the United States government
Recently, Deep Image Prior (DIP) has emerged as an effective unsupervised one-shot learner, delivering competitive results across various image recovery problems. This method only requires the noisy measurements and a forward operator, relying solely on deep networks initialized with random noise to learn and restore the structure of the data. However, DIP is notorious for its vulnerability to overfitting due to the overparameterization of the network. Building upon insights into the impact of the DIP input and drawing inspiration from the gradual denoising process in cutting-edge diffusion models, we introduce Autoencoding Sequential DIP (aSeqDIP) for image reconstruction. This method progressively denoises and reconstructs the image through a sequential optimization of network weights. This is achieved using an input-adaptive DIP objective, combined with an autoencoding regularization term. Compared to diffusion models, our method does not require training data and outperforms other DIP-based methods in mitigating noise overfitting while maintaining a similar number of parameter updates as Vanilla DIP. Through extensive experiments, we validate the effectiveness of our method in various image reconstruction tasks, such as MRI and CT reconstruction, as well as in image restoration tasks like image denoising, inpainting, and non-linear deblurring.
more »
« less
This content will become publicly available on July 6, 2027
Normalization-equivariant Diffusion Models: Learning Posterior Samplers From Noisy And Partial Measurements,
Diffusion models (DMs) are a powerful frame- work for image generation and restoration. How- ever, existing DMs are primarily trained in a su- pervised manner by using a large corpus of clean images. This poses fundamental challenges in many real-world scenarios, where acquiring noise- free data is hard or infeasible. While some meth- ods are capable of training DMs using noisy data, they are effective only when the amount of noise is very mild or when additional noise-free data is available. In addition, existing methods for training DMs from incomplete measurements re- quire access to multiple complementary acquisi- tion processes, a significant practical limitation. Here we introduce the first approach for learning DMs for image restoration using only noisy mea- surement data from a single operator. First, we show that DMs, and more broadly minimum mean squared error denoisers, exhibit a weak form of scale equivariance linking rescaling in signal am- plitude to changes in noise intensity. We then leverage this theoretical insight to develop a de- noising score-matching strategy that generalizes robustly to noise levels below the training data, thereby enabling the learning of DMs from noisy measurements. For problems involving measure- ments both noisy and incomplete, we integrate our method with equivariant imaging, a complemen- tary self-supervised learning framework that ex- ploits the inherent invariants of imaging problems. This allows training DMs for image restoration from single-operator noisy measurements. We validate the effectiveness of our approach through extensive experiments on image denoising, demo- saicing, inpainting, and MRI reconstruction along with comparisons with the state of the art.
more »
« less
- Award ID(s):
- 2239687
- PAR ID:
- 10696767
- Publisher / Repository:
- International Conference on Machine Learning
- Date Published:
- Format(s):
- Medium: X
- Location:
- Seoul, Korea
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Image restoration aims to recover a clean image given a noisy image. It has long been a topic of interest for researchers in imaging, optical science and computer vision. As the imaging environment becomes more and more deteriorated, the problem becomes more challenging. Several computational approaches, ranging from statistical to deep learning, have been proposed over the years to tackle this problem. The deep learning-based approaches provided promising image restoration results, but it’s purely data driven and the requirement of large datasets (paired or unpaired) for training might demean its utility for certain physical problems. Recently, physics informed image restoration techniques have gained importance due to their ability to enhance performance, infer some sense of the degradation process and its potential to quantify the uncertainty in the prediction results. In this paper, we propose a physics informed deep learning approach with simultaneous parameter estimation using 3D integral imaging and Bayesian neural network (BNN). An image-image mapping architecture is first pretrained to generate a clean image from the degraded image, which is then utilized for simultaneous training with Bayesian neural network for simultaneous parameter estimation. For the network training, simulated data using the physical model has been utilized instead of actual degraded data. The proposed approach has been tested experimentally under degradations such as low illumination and partial occlusion. The recovery results are promising despite training from a simulated dataset. We have tested the performance of the approach under varying levels of illumination condition. Additionally, the proposed approach also has been analyzed against corresponding 2D imaging-based approach. The results suggest significant improvements compared to 2D even training under similar datasets. Also, the parameter estimation results demonstrate the utility of the approach in estimating the degradation parameter in addition to image restoration under the experimental conditions considered.more » « less
-
null (Ed.)Deep neural networks have emerged as very successful tools for image restoration and reconstruction tasks. These networks are often trained end-to-end to directly reconstruct an image from a noisy or corrupted measurement of that image. To achieve state-of-the-art performance, training on large and diverse sets of images is considered critical. However, it is often difficult and/or expensive to collect large amounts of training images. Inspired by the success of Data Augmentation (DA) for classification problems, in this paper, we propose a pipeline for data augmentation for accelerated MRI reconstruction and study its effectiveness at reducing the required training data in a variety of settings. Our DA pipeline, MRAugment, is specifically designed to utilize the invariances present in medical imaging measurements as naive DA strategies that neglect the physics of the problem fail. Through extensive studies on multiple datasets we demonstrate that in the low-data regime DA prevents overfitting and can match or even surpass the state of the art while using significantly fewer training data, whereas in the high-data regime it has diminishing returns. Furthermore, our findings show that DA improves the robustness of the model against various shifts in the test distribution.more » « less
-
Deep neural networks have emerged as very successful tools for image restoration and reconstruction tasks. These networks are often trained end-to-end to directly reconstruct an image from a noisy or corrupted measurement of that image. To achieve state-of-the-art performance, training on large and diverse sets of images is considered critical. However, it is often difficult and/or expensive to collect large amounts of training images. Inspired by the success of Data Augmentation (DA) for classification problems, in this paper, we propose a pipeline for data augmentation for accelerated MRI reconstruction and study its effectiveness at reducing the required training data in a variety of settings. Our DA pipeline, MRAugment, is specifically designed to utilize the invariances present in medical imaging measurements as naive DA strategies that neglect the physics of the problem fail. Through extensive studies on multiple datasets we demonstrate that in the low-data regime DA prevents overfitting and can match or even surpass the state of the art while using significantly fewer training data, whereas in the high-data regime it has diminishing returns. Furthermore, our findings show that DA can improve the robustness of the model against various shifts in the test distribution.more » « less
-
Federated learning (FL) is a learning paradigm that allows the central server to learn from different data sources while keeping the data private locally. Without controlling and monitoring the local data collection process, the locally available training labels are likely noisy, i.e., the collected training labels differ from the unobservable ground truth. Additionally, in heterogenous FL, each local client may only have access to a subset of label space (referred to as openset label learning), meanwhile without overlapping with others. In this work, we study the challenge of FL with local openset noisy labels. We observe that many existing solutions in the noisy label literature, e.g., loss correction, are ineffective during local training due to overfitting to noisy labels and being not generalizable to openset labels. For the methods in FL, different estimated metrics are shared. To address the problems, we design a label communication mechanism that shares "contrastive labels" randomly selected from clients with the server. The privacy of the shared contrastive labels is protected by label differential privacy (DP). Both the DP guarantee and the effectiveness of our approach are theoretically guaranteed. Compared with several baseline methods, our solution shows its efficiency in several public benchmarks and real-world datasets under different noise ratios and noise models.more » « less
