Sparse-view computed tomography (CT) enables fast and low-dose CT imaging, an essential feature for patient-save medical imaging and rapid non-destructive testing. In sparse-view CT, only a few projection views are acquired, causing standard reconstructions to suffer from severe artifacts and noise. To address these issues, we propose a self-supervised image reconstruction strategy. Specifically, in contrast to the established Noise2Inverse, our proposed training strategy uses a loss function in the projection domain, thereby bypassing the otherwise prescribed nullspace component. We demonstrate the effectiveness of the proposed method in reducing stripe-artifacts and noise, even from highly sparse data.
Neural Radiance Fields (NeRF), as a pioneering technique in computer vision, offer great potential to revolutionize medical imaging by synthesizing three-dimensional representations from the projected two-dimensional image data. However, they face unique challenges when applied to medical applications. This paper presents a comprehensive examination of applications of NeRFs in medical imaging, highlighting four imminent challenges, including fundamental imaging principles, inner structure requirement, object boundary definition, and color density significance. We discuss current methods on different organs and discuss related limitations. We also review several datasets and evaluation metrics and propose several promising directions for future research.
The quality of the data and annotation upper-bounds the quality of a downstream model. While there exist large text corpora and image-text pairs, high-quality video-text data is much harder to collect. First of all, manual labeling is more time-consuming, as it requires an annotator to watch an entire video. Second, videos have a temporal dimension, consisting of several scenes stacked together, and showing multiple actions. Accordingly, to establish a video dataset with high-quality captions, we propose an automatic approach leveraging multimodal inputs, such as textual video description, subtitles, and individual video frames. Specifically, we curate 3.8M high-resolution videos from the publicly available HD-VILA-100M dataset. We then split them into semantically consistent video clips, and apply multiple cross-modality teacher models to obtain captions for each video. Next, we finetune a retrieval model on a small subset where the best caption of each video is manually selected and then employ the model in the whole dataset to select the best caption as the annotation. In this way, we get 70M videos paired with high-quality text captions. We dub the dataset as Panda-70M. We show the value of the proposed dataset on three downstream tasks: video captioning, video and text retrieval, and text-driven video generation. The models trained on the proposed data score substantially better on the majority of metrics across all the tasks.
Inverse problems span across diverse fields. In medical contexts, computed tomography (CT) plays a crucial role in reconstructing a patient's internal structure, presenting challenges due to artifacts caused by inherently ill-posed inverse problems. Previous research advanced image quality via post-processing and deep unrolling algorithms but faces challenges, such as extended convergence times with ultra-sparse data. Despite enhancements, resulting images often show significant artifacts, limiting their effectiveness for real-world diagnostic applications. We aim to explore deep second-order unrolling algorithms for solving imaging inverse problems, emphasizing their faster convergence and lower time complexity compared to common first-order methods like gradient descent. In this paper, we introduce QN-Mixer, an algorithm based on the quasi-Newton approach. We use learned parameters through the BFGS algorithm and introduce Incept-Mixer, an efficient neural architecture that serves as a non-local regularization term, capturing long-range dependencies within images. To address the computational demands typically associated with quasi-Newton algorithms that require full Hessian matrix computations, we present a memory-efficient alternative. Our approach intelligently downsamples gradient information, significantly reducing computational requirements while maintaining performance. The approach is validated through experiments on the sparse-view CT problem, involving various datasets and scanning protocols, and is compared with post-processing and deep unrolling state-of-the-art approaches. Our method outperforms existing approaches and achieves state-of-the-art performance in terms of SSIM and PSNR, all while reducing the number of unrolling iterations required.
The reconstruction of 3D microstructures from 2D slices is considered to hold significant value in predicting the spatial structure and physical properties of materials.The dimensional extension from 2D to 3D is viewed as a highly challenging inverse problem from the current technological perspective.Recently,methods based on generative adversarial networks have garnered widespread attention.However,they are still hampered by numerous limitations,including oversimplified models,a requirement for a substantial number of training samples,and difficulties in achieving model convergence during training.In light of this,a novel generative model that integrates the multiscale properties of U-net with and the generative capabilities of GAN has been proposed.Based on this,the innovative construction of a multi-scale channel aggregation module,a multi-scale hierarchical feature aggregation module and a convolutional block attention mechanism can better capture the properties of the material microstructure and extract the image information.The model's accuracy is further improved by combining the image regularization loss with the Wasserstein distance loss.In addition,this study utilizes the anisotropy index to accurately distinguish the nature of the image,which can clearly determine the isotropy and anisotropy of the image.It is also the first time that the generation quality of material samples from different domains is evaluated and the performance of the model itself is compared.The experimental results demonstrate that the present model not only shows a very high similarity between the generated 3D structures and real samples but is also highly consistent with real data in terms of statistical data analysis.
Large occlusions result in a significant decline in image classification accuracy. During inference, diverse types of unseen occlusions introduce out-of-distribution data to the classification model, leading to accuracy dropping as low as 50%. As occlusions encompass spatially connected regions, conventional methods involving feature reconstruction are inadequate for enhancing classification performance. We introduce LEARN: Latent Enhancing feAture Reconstruction Network -- An auto-encoder based network that can be incorporated into the classification model before its classifier head without modifying the weights of classification model. In addition to reconstruction and classification losses, training of LEARN effectively combines intra- and inter-class losses calculated over its latent space -- which lead to improvement in recovering latent space of occluded data, while preserving its class-specific discriminative information. On the OccludedPASCAL3D+ dataset, the proposed LEARN outperforms standard classification models (VGG16 and ResNet-50) by a large margin and up to 2% over state-of-the-art methods. In cross-dataset testing, our method improves the average classification accuracy by more than 5% over the state-of-the-art methods. In every experiment, our model consistently maintains excellent accuracy on in-distribution data.
Diffusion and Poisson flow models have demonstrated remarkable success for a wide range of generative tasks. Nevertheless, their iterative nature results in computationally expensive sampling and the number of function evaluations (NFE) required can be orders of magnitude larger than for single-step methods. Consistency models are a recent class of deep generative models which enable single-step sampling of high quality data without the need for adversarial training. In this paper, we introduce a novel image denoising technique which combines the flexibility afforded in Poisson flow generative models (PFGM)++ with the, high quality, single step sampling of consistency models. The proposed method first learns a trajectory between a noise distribution and the posterior distribution of interest by training PFGM++ in a supervised fashion. These pre-trained PFGM++ are subsequently "distilled" into Poisson flow consistency models (PFCM) via an updated version of consistency distillation. We call this approach posterior sampling Poisson flow consistency models (PS-PFCM). Our results indicate that the added flexibility of tuning the hyperparameter D, the dimensionality of the augmentation variables in PFGM++, allows us to outperform consistency models, a current state-of-the-art diffusion-style model with NFE=1 on clinical low-dose CT images. Notably, PFCM is in itself a novel family of deep generative models and we provide initial results on the CIFAR-10 dataset.
Diabetes mellitus (DM) predisposes patients to vascular complications. Retinal images and vasculature reflect the body's micro- and macrovascular health. They can be used to diagnose DM complications, including diabetic retinopathy (DR), neuropathy, nephropathy, and atherosclerotic cardiovascular disease, as well as forecast the risk of cardiovascular events. Artificial intelligence (AI)-enabled systems developed for high-throughput detection of DR using digitized retinal images have become clinically adopted. Beyond DR screening, AI integration also holds immense potential to address challenges associated with the holistic care of the patient with DM. In this work, we aim to comprehensively review the literature for studies on AI applications based on retinal images related to DM diagnosis, prognostication, and management. We will describe the findings of holistic AI-assisted diabetes care, including but not limited to DR screening, and discuss barriers to implementing such systems, including issues concerning ethics, data privacy, equitable access, and explainability. With the ability to evaluate the patient's health status vis a vis DM complication as well as risk prognostication of future cardiovascular complications, AI-assisted retinal image analysis has the potential to become a central tool for modern personalized medicine in patients with DM.
Deep superpixel algorithms have made remarkable strides by substituting hand-crafted features with learnable ones. Nevertheless, we observe that existing deep superpixel methods, serving as mid-level representation operations, remain sensitive to the statistical properties (e.g., color distribution, high-level semantics) embedded within the training dataset. Consequently, learnable features exhibit constrained discriminative capability, resulting in unsatisfactory pixel grouping performance, particularly in untrainable application scenarios. To address this issue, we propose the Content Disentangle Superpixel (CDS) algorithm to selectively separate the invariant inter-pixel correlations and statistical properties, i.e., style noise. Specifically, We first construct auxiliary modalities that are homologous to the original RGB image but have substantial stylistic variations. Then, driven by mutual information, we propose the local-grid correlation alignment across modalities to reduce the distribution discrepancy of adaptively selected features and learn invariant inter-pixel correlations. Afterwards, we perform global-style mutual information minimization to enforce the separation of invariant content and train data styles. The experimental results on four benchmark datasets demonstrate the superiority of our approach to existing state-of-the-art methods, regarding boundary adherence, generalization, and efficiency. Code and pre-trained model are available at https://github.com/rookiie/CDSpixel.
This paper proposes a method for Acoustic Constrained Segmentation (ACS) in audio recordings of vehicles driven through a production test track, delimiting the boundaries of surface types in the track. ACS is a variant of classical acoustic segmentation where the sequence of labels is known, contiguous and invariable, which is especially useful in this work as the test track has a standard configuration of surface types. The proposed ConvDTW-ACS method utilizes a Convolutional Neural Network for classifying overlapping image chunks extracted from the full audio spectrogram. Then, our custom Dynamic Time Warping algorithm aligns the sequence of predicted probabilities to the sequence of surface types in the track, from which timestamps of the surface type boundaries can be extracted. The method was evaluated on a real-world dataset collected from the Ford Manufacturing Plant in Valencia (Spain), achieving a mean error of 166 milliseconds when delimiting, within the audio, the boundaries of the surfaces in the track. The results demonstrate the effectiveness of the proposed method in accurately segmenting different surface types, which could enable the development of more specialized AI systems to improve the quality inspection process.