The enhanced performance of AI has accelerated its integration into scientific research. In particular, the use of generative AI to create scientific hypotheses is promising and is increasingly being applied across various fields. However, when employing AI-generated hypotheses for critical decisions, such as medical diagnoses, verifying their reliability is crucial. In this study, we consider a medical diagnostic task using generated images by diffusion models, and propose a statistical test to quantify its reliability. The basic idea behind the proposed statistical test is to employ a selective inference framework, where we consider a statistical test conditional on the fact that the generated images are produced by a trained diffusion model. Using the proposed method, the statistical reliability of medical image diagnostic results can be quantified in the form of a p-value, allowing for decision-making with a controlled error rate. We show the theoretical validity of the proposed statistical test and its effectiveness through numerical experiments on synthetic and brain image datasets.
In recent years, machine learning has witnessed extensive adoption across various sectors, yet its application in medical image-based disease detection and diagnosis remains challenging due to distribution shifts in real-world data. In practical settings, deployed models encounter samples that differ significantly from the training dataset, especially in the health domain, leading to potential performance issues. This limitation hinders the expressiveness and reliability of deep learning models in health applications. Thus, it becomes crucial to identify methods capable of producing reliable uncertainty estimation in the context of distribution shifts in the health sector. In this paper, we explore the feasibility of using cutting-edge Bayesian and non-Bayesian methods to detect distributionally shifted samples, aiming to achieve reliable and trustworthy diagnostic predictions in segmentation task. Specifically, we compare three distinct uncertainty estimation methods, each designed to capture either unimodal or multimodal aspects in the posterior distribution. Our findings demonstrate that methods capable of addressing multimodal characteristics in the posterior distribution, offer more dependable uncertainty estimates. This research contributes to enhancing the utility of deep learning in healthcare, making diagnostic predictions more robust and trustworthy.
In the field of biomedical image analysis, the quest for architectures capable of effectively capturing long-range dependencies is paramount, especially when dealing with 3D image segmentation, classification, and landmark detection. Traditional Convolutional Neural Networks (CNNs) struggle with locality respective field, and Transformers have a heavy computational load when applied to high-dimensional medical images. In this paper, we introduce nnMamba, a novel architecture that integrates the strengths of CNNs and the advanced long-range modeling capabilities of State Space Sequence Models (SSMs). nnMamba adds the SSMs to the convolutional residual-block to extract local features and model complex dependencies. For diffirent tasks, we build different blocks to learn the features. Extensive experiments demonstrate nnMamba's superiority over state-of-the-art methods in a suite of challenging tasks, including 3D image segmentation, classification, and landmark detection. nnMamba emerges as a robust solution, offering both the local representation ability of CNNs and the efficient global context processing of SSMs, setting a new standard for long-range dependency modeling in medical image analysis. Code is available at https://github.com/lhaof/nnMamba
Emerging Learned image Compression (LC) achieves significant improvements in coding efficiency by end-to-end training of neural networks for compression. An important benefit of this approach over traditional codecs is that any optimization criteria can be directly applied to the encoder-decoder networks during training. Perceptual optimization of LC to comply with the Human Visual System (HVS) is among such criteria, which has not been fully explored yet. This paper addresses this gap by proposing a novel framework to integrate Just Noticeable Distortion (JND) principles into LC. Leveraging existing JND datasets, three perceptual optimization methods are proposed to integrate JND into the LC training process: (1) Pixel-Wise JND Loss (PWL) prioritizes pixel-by-pixel fidelity in reproducing JND characteristics, (2) Image-Wise JND Loss (IWL) emphasizes on overall imperceptible degradation levels, and (3) Feature-Wise JND Loss (FWL) aligns the reconstructed image features with perceptually significant features. Experimental evaluations demonstrate the effectiveness of JND integration, highlighting improvements in rate-distortion performance and visual quality, compared to baseline methods. The proposed methods add no extra complexity after training.
Visual Question Answering (VQA) is a fundamental task in computer vision and natural language process fields. Although the ``pre-training & finetuning'' learning paradigm significantly improves the VQA performance, the adversarial robustness of such a learning paradigm has not been explored. In this paper, we delve into a new problem: using a pre-trained multimodal source model to create adversarial image-text pairs and then transferring them to attack the target VQA models. Correspondingly, we propose a novel VQAttack model, which can iteratively generate both image and text perturbations with the designed modules: the large language model (LLM)-enhanced image attack and the cross-modal joint attack module. At each iteration, the LLM-enhanced image attack module first optimizes the latent representation-based loss to generate feature-level image perturbations. Then it incorporates an LLM to further enhance the image perturbations by optimizing the designed masked answer anti-recovery loss. The cross-modal joint attack module will be triggered at a specific iteration, which updates the image and text perturbations sequentially. Notably, the text perturbation updates are based on both the learned gradients in the word embedding space and word synonym-based substitution. Experimental results on two VQA datasets with five validated models demonstrate the effectiveness of the proposed VQAttack in the transferable attack setting, compared with state-of-the-art baselines. This work reveals a significant blind spot in the ``pre-training & fine-tuning'' paradigm on VQA tasks. Source codes will be released.
The Transiting Exoplanet Survey Satellite (TESS) mission measured light from stars in ~85% of the sky throughout its two-year primary mission, resulting in millions of TESS 30-minute cadence light curves to analyze in the search for transiting exoplanets. To search this vast dataset, we aim to provide an approach that is both computationally efficient, produces highly performant predictions, and minimizes the required human search effort. We present a convolutional neural network that we train to identify short period variables. To make a prediction for a given light curve, our network requires no prior target parameters identified using other methods. Our network performs inference on a TESS 30-minute cadence light curve in ~5ms on a single GPU, enabling large scale archival searches. We present a collection of 14156 short-period variables identified by our network. The majority of our identified variables fall into two prominent populations, one of short-period main sequence binaries and another of Delta Scuti stars. Our neural network model and related code is additionally provided as open-source code for public use and extension.
The newly released Segment Anything Model (SAM) is a popular tool used in image processing due to its superior segmentation accuracy, variety of input prompts, training capabilities, and efficient model design. However, its current model is trained on a diverse dataset not tailored to medical images, particularly ultrasound images. Ultrasound images tend to have a lot of noise, making it difficult to segment out important structures. In this project, we developed ClickSAM, which fine-tunes the Segment Anything Model using click prompts for ultrasound images. ClickSAM has two stages of training: the first stage is trained on single-click prompts centered in the ground-truth contours, and the second stage focuses on improving the model performance through additional positive and negative click prompts. By comparing the first stage predictions to the ground-truth masks, true positive, false positive, and false negative segments are calculated. Positive clicks are generated using the true positive and false negative segments, and negative clicks are generated using the false positive segments. The Centroidal Voronoi Tessellation algorithm is then employed to collect positive and negative click prompts in each segment that are used to enhance the model performance during the second stage of training. With click-train methods, ClickSAM exhibits superior performance compared to other existing models for ultrasound image segmentation.
Taming the generation outcome of state of the art Diffusion and Flow-Matching (FM) models without having to re-train a task-specific model unlocks a powerful tool for solving inverse problems, conditional generation, and controlled generation in general. In this work we introduce D-Flow, a simple framework for controlling the generation process by differentiating through the flow, optimizing for the source (noise) point. We motivate this framework by our key observation stating that for Diffusion/FM models trained with Gaussian probability paths, differentiating through the generation process projects gradient on the data manifold, implicitly injecting the prior into the optimization process. We validate our framework on linear and non-linear controlled generation problems including: image and audio inverse problems and conditional molecule generation reaching state of the art performance across all.
Camera relocalization is pivotal in computer vision, with applications in AR, drones, robotics, and autonomous driving. It estimates 3D camera position and orientation (6-DoF) from images. Unlike traditional methods like SLAM, recent strides use deep learning for direct end-to-end pose estimation. We propose EffLoc, a novel efficient Vision Transformer for single-image camera relocalization. EffLoc's hierarchical layout, memory-bound self-attention, and feed-forward layers boost memory efficiency and inter-channel communication. Our introduced sequential group attention (SGA) module enhances computational efficiency by diversifying input features, reducing redundancy, and expanding model capacity. EffLoc excels in efficiency and accuracy, outperforming prior methods, such as AtLoc and MapNet. It thrives on large-scale outdoor car-driving scenario, ensuring simplicity, end-to-end trainability, and eliminating handcrafted loss functions.
The process of camera calibration involves estimating the intrinsic and extrinsic parameters, which are essential for accurately performing tasks such as 3D reconstruction, object tracking and augmented reality. In this work, we propose a novel constraints-based loss for measuring the intrinsic (focal length: $(f_x, f_y)$ and principal point: $(p_x, p_y)$) and extrinsic (baseline: ($b$), disparity: ($d$), translation: $(t_x, t_y, t_z)$, and rotation specifically pitch: $(\theta_p)$) camera parameters. Our novel constraints are based on geometric properties inherent in the camera model, including the anatomy of the projection matrix (vanishing points, image of world origin, axis planes) and the orthonormality of the rotation matrix. Thus we proposed a novel Unsupervised Geometric Constraint Loss (UGCL) via a multitask learning framework. Our methodology is a hybrid approach that employs the learning power of a neural network to estimate the desired parameters along with the underlying mathematical properties inherent in the camera projection matrix. This distinctive approach not only enhances the interpretability of the model but also facilitates a more informed learning process. Additionally, we introduce a new CVGL Camera Calibration dataset, featuring over 900 configurations of camera parameters, incorporating 63,600 image pairs that closely mirror real-world conditions. By training and testing on both synthetic and real-world datasets, our proposed approach demonstrates improvements across all parameters when compared to the state-of-the-art (SOTA) benchmarks. The code and the updated dataset can be found here: https://github.com/CVLABLUMS/CVGL-Camera-Calibration