Transfer learning is a promising technique for medical image classification, particularly for long-tailed datasets. However, the scarcity of data in medical imaging domains often leads to overparameterization when fine-tuning large publicly available pre-trained models. Moreover, these large models are ineffective in deployment in clinical settings due to their computational expenses. To address these challenges, we propose FoPro-KD, a novel approach that unleashes the power of frequency patterns learned from frozen publicly available pre-trained models to enhance their transferability and compression. FoPro-KD comprises three modules: Fourier prompt generator (FPG), effective knowledge distillation (EKD), and adversarial knowledge distillation (AKD). The FPG module learns to generate targeted perturbations conditional on a target dataset, exploring the representations of a frozen pre-trained model, trained on natural images. The EKD module exploits these generalizable representations through distillation to a smaller target model, while the AKD module further enhances the distillation process. Through these modules, FoPro-KD achieves significant improvements in performance on long-tailed medical image classification benchmarks, demonstrating the potential of leveraging the learned frequency patterns from pre-trained models to enhance transfer learning and compression of large pre-trained models for feasible deployment.
Medical image data are often limited due to the expensive acquisition and annotation process. Hence, training a deep-learning model with only raw data can easily lead to overfitting. One solution to this problem is to augment the raw data with various transformations, improving the model's ability to generalize to new data. However, manually configuring a generic augmentation combination and parameters for different datasets is non-trivial due to inconsistent acquisition approaches and data distributions. Therefore, automatic data augmentation is proposed to learn favorable augmentation strategies for different datasets while incurring large GPU overhead. To this end, we present a novel method, called Dynamic Data Augmentation (DDAug), which is efficient and has negligible computation cost. Our DDAug develops a hierarchical tree structure to represent various augmentations and utilizes an efficient Monte-Carlo tree searching algorithm to update, prune, and sample the tree. As a result, the augmentation pipeline can be optimized for each dataset automatically. Experiments on multiple Prostate MRI datasets show that our method outperforms the current state-of-the-art data augmentation strategies.
This study investigates barely-supervised medical image segmentation where only few labeled data, i.e., single-digit cases are available. We observe the key limitation of the existing state-of-the-art semi-supervised solution cross pseudo supervision is the unsatisfactory precision of foreground classes, leading to a degenerated result under barely-supervised learning. In this paper, we propose a novel Compete-to-Win method (ComWin) to enhance the pseudo label quality. In contrast to directly using one model's predictions as pseudo labels, our key idea is that high-quality pseudo labels should be generated by comparing multiple confidence maps produced by different networks to select the most confident one (a compete-to-win strategy). To further refine pseudo labels at near-boundary areas, an enhanced version of ComWin, namely, ComWin+, is proposed by integrating a boundary-aware enhancement module. Experiments show that our method can achieve the best performance on three public medical image datasets for cardiac structure segmentation, pancreas segmentation and colon tumor segmentation, respectively. The source code is now available at https://github.com/Huiimin5/comwin.
Contrastive Language-Image Pre-training (CLIP) is a powerful multimodal large vision model that has demonstrated significant benefits for downstream tasks, including many zero-shot learning and text-guided vision tasks. However, we notice some severe problems regarding the model's explainability, which undermines its credibility and impedes related tasks. Specifically, we find CLIP prefers the background regions than the foregrounds according to the predicted similarity map, which contradicts human understanding. Besides, there are obvious noisy activations on the visualization results at irrelevant positions. To address these two issues, we conduct in-depth analyses and reveal the reasons with new findings and evidences. Based on these insights, we propose the CLIP Surgery, a method that enables surgery-like modifications for the inference architecture and features, for better explainability and enhancement in multiple open-vocabulary tasks. The proposed method has significantly improved the explainability of CLIP for both convolutional networks and vision transformers, surpassing existing methods by large margins. Besides, our approach also demonstrates remarkable improvements in open-vocabulary segmentation and multi-label recognition tasks. For examples, the mAP improvement on NUS-Wide multi-label recognition is 4.41% without any additional training, and our CLIP Surgery surpasses the state-of-the-art method by 8.74% at mIoU on Cityscapes open-vocabulary semantic segmentation. Furthermore, our method benefits other tasks including multimodal visualization and interactive segmentation like Segment Anything Model (SAM). The code is available at https://github.com/xmed-lab/CLIP_Surgery
Deep hashing has been extensively applied to massive image retrieval due to its efficiency and effectiveness. Recently, several adversarial attacks have been presented to reveal the vulnerability of deep hashing models against adversarial examples. However, existing attack methods suffer from degraded performance or inefficiency because they underutilize the semantic relations between original samples or spend a lot of time learning these relations with a deep neural network. In this paper, we propose a novel Pharos-guided Attack, dubbed PgA, to evaluate the adversarial robustness of deep hashing networks reliably and efficiently. Specifically, we design pharos code to represent the semantics of the benign image, which preserves the similarity to semantically relevant samples and dissimilarity to irrelevant ones. It is proven that we can quickly calculate the pharos code via a simple math formula. Accordingly, PgA can directly conduct a reliable and efficient attack on deep hashing-based retrieval by maximizing the similarity between the hash code of the adversarial example and the pharos code. Extensive experiments on the benchmark datasets verify that the proposed algorithm outperforms the prior state-of-the-arts in both attack strength and speed.
Optical Coherence Tomography Angiography (OCTA) has become increasingly vital in the clinical screening of fundus diseases due to its ability to capture accurate 3D imaging of blood vessels in a non-contact scanning manner. However, the acquisition of OCTA images remains challenging due to the requirement of exclusive sensors and expensive devices. In this paper, we propose a novel framework, TransPro, that translates 3D Optical Coherence Tomography (OCT) images into exclusive 3D OCTA images using an image translation pattern. Our main objective is to address two issues in existing image translation baselines, namely, the aimlessness in the translation process and incompleteness of the translated object. The former refers to the overall quality of the translated OCTA images being satisfactory, but the retinal vascular quality being low. The latter refers to incomplete objects in translated OCTA images due to the lack of global contexts. TransPro merges a 2D retinal vascular segmentation model and a 2D OCTA image translation model into a 3D image translation baseline for the 2D projection map projected by the translated OCTA images. The 2D retinal vascular segmentation model can improve attention to the retinal vascular, while the 2D OCTA image translation model introduces beneficial heuristic contextual information. Extensive experimental results on two challenging datasets demonstrate that TransPro can consistently outperform existing approaches with minimal computational overhead during training and none during testing.
Sparse-view cone-beam CT (CBCT) reconstruction is an important direction to reduce radiation dose and benefit clinical applications. Previous voxel-based generation methods represent the CT as discrete voxels, resulting in high memory requirements and limited spatial resolution due to the use of 3D decoders. In this paper, we formulate the CT volume as a continuous intensity field and develop a novel DIF-Net to perform high-quality CBCT reconstruction from extremely sparse (fewer than 10) projection views at an ultrafast speed. The intensity field of a CT can be regarded as a continuous function of 3D spatial points. Therefore, the reconstruction can be reformulated as regressing the intensity value of an arbitrary 3D point from given sparse projections. Specifically, for a point, DIF-Net extracts its view-specific features from different 2D projection views. These features are subsequently aggregated by a fusion module for intensity estimation. Notably, thousands of points can be processed in parallel to improve efficiency during training and testing. In practice, we collect a knee CBCT dataset to train and evaluate DIF-Net. Extensive experiments show that our approach can reconstruct CBCT with high image quality and high spatial resolution from extremely sparse views within 1.6 seconds, significantly outperforming state-of-the-art methods. Our code will be available at https://github.com/lyqun/DIF-Net.
Deep regression is an important problem with numerous applications. These range from computer vision tasks such as age estimation from photographs, to medical tasks such as ejection fraction estimation from echocardiograms for disease tracking. Semi-supervised approaches for deep regression are notably under-explored compared to classification and segmentation tasks, however. Unlike classification tasks, which rely on thresholding functions for generating class pseudo-labels, regression tasks use real number target predictions directly as pseudo-labels, making them more sensitive to prediction quality. In this work, we propose a novel approach to semi-supervised regression, namely Uncertainty-Consistent Variational Model Ensembling (UCVME), which improves training by generating high-quality pseudo-labels and uncertainty estimates for heteroscedastic regression. Given that aleatoric uncertainty is only dependent on input data by definition and should be equal for the same inputs, we present a novel uncertainty consistency loss for co-trained models. Our consistency loss significantly improves uncertainty estimates and allows higher quality pseudo-labels to be assigned greater importance under heteroscedastic regression. Furthermore, we introduce a novel variational model ensembling approach to reduce prediction noise and generate more robust pseudo-labels. We analytically show our method generates higher quality targets for unlabeled data and further improves training. Experiments show that our method outperforms state-of-the-art alternatives on different tasks and can be competitive with supervised methods that use full labels. Our code is available at https://github.com/xmed-lab/UCVME.
Cardiovascular disease (CVD) accounts for about half of non-communicable diseases. Vessel stenosis in the coronary artery is considered to be the major risk of CVD. Computed tomography angiography (CTA) is one of the widely used noninvasive imaging modalities in coronary artery diagnosis due to its superior image resolution. Clinically, segmentation of coronary arteries is essential for the diagnosis and quantification of coronary artery disease. Recently, a variety of works have been proposed to address this problem. However, on one hand, most works rely on in-house datasets, and only a few works published their datasets to the public which only contain tens of images. On the other hand, their source code have not been published, and most follow-up works have not made comparison with existing works, which makes it difficult to judge the effectiveness of the methods and hinders the further exploration of this challenging yet critical problem in the community. In this paper, we propose a large-scale dataset for coronary artery segmentation on CTA images. In addition, we have implemented a benchmark in which we have tried our best to implement several typical existing methods. Furthermore, we propose a strong baseline method which combines multi-scale patch fusion and two-stage processing to extract the details of vessels. Comprehensive experiments show that the proposed method achieves better performance than existing works on the proposed large-scale dataset. The benchmark and the dataset are published at https://github.com/XiaoweiXu/ImageCAS-A-Large-Scale-Dataset-and-Benchmark-for-Coronary-Artery-Segmentation-based-on-CT.
Left-ventricular ejection fraction (LVEF) is an important indicator of heart failure. Existing methods for LVEF estimation from video require large amounts of annotated data to achieve high performance, e.g. using 10,030 labeled echocardiogram videos to achieve mean absolute error (MAE) of 4.10. Labeling these videos is time-consuming however and limits potential downstream applications to other heart diseases. This paper presents the first semi-supervised approach for LVEF prediction. Unlike general video prediction tasks, LVEF prediction is specifically related to changes in the left ventricle (LV) in echocardiogram videos. By incorporating knowledge learned from predicting LV segmentations into LVEF regression, we can provide additional context to the model for better predictions. To this end, we propose a novel Cyclical Self-Supervision (CSS) method for learning video-based LV segmentation, which is motivated by the observation that the heartbeat is a cyclical process with temporal repetition. Prediction masks from our segmentation model can then be used as additional input for LVEF regression to provide spatial context for the LV region. We also introduce teacher-student distillation to distill the information from LV segmentation masks into an end-to-end LVEF regression model that only requires video inputs. Results show our method outperforms alternative semi-supervised methods and can achieve MAE of 4.17, which is competitive with state-of-the-art supervised performance, using half the number of labels. Validation on an external dataset also shows improved generalization ability from using our method.