Rabab Ward

Li Guo, Anas M. Tahir, Dong Zhang et al.

The increasing demand for medical imaging has surpassed the capacity of available radiologists, leading to diagnostic delays and potential misdiagnoses. Artificial intelligence (AI) techniques, particularly in automatic medical report generation (AMRG), offer a promising solution to this dilemma. This review comprehensively examines AMRG methods from 2021 to 2024. It (i) presents solutions to primary challenges in this field, (ii) explores AMRG applications across various imaging modalities, (iii) introduces publicly available datasets, (iv) outlines evaluation metrics, (v) identifies techniques that significantly enhance model performance, and (vi) discusses unresolved issues and potential future research directions. This paper aims to provide a comprehensive understanding of the existing literature and inspire valuable future research.

Mazen Abdelfattah, Kaiwen Yuan, Z. Jane Wang et al.

Modern autonomous vehicles rely heavily on mechanical LiDARs for perception. Current perception methods generally require 360° point clouds, collected sequentially as the LiDAR scans the azimuth and acquires consecutive wedge-shaped slices. The acquisition latency of a full scan (~ 100ms) may lead to outdated perception which is detrimental to safe operation. Recent streaming perception works proposed directly processing LiDAR slices and compensating for the narrow field of view (FOV) of a slice by reusing features from preceding slices. These works, however, are all based on a single modality and require past information which may be outdated. Meanwhile, images from high-frequency cameras can support streaming models as they provide a larger FoV compared to a LiDAR slice. However, this difference in FoV complicates sensor fusion. To address this research gap, we propose an innovative camera-LiDAR streaming 3D object detection framework that uses camera images instead of past LiDAR slices to provide an up-to-date, dense, and wide context for streaming perception. The proposed method outperforms prior streaming models on the challenging NuScenes benchmark. It also outperforms powerful full-scan detectors while being much faster. Our method is shown to be robust to missing camera images, narrow LiDAR slices, and small camera-LiDAR miscalibration.

Peizhi Yan, Rabab Ward, Qiang Tang et al.

Recent advancements in 3D Gaussian Splatting (3DGS) have unlocked significant potential for modeling 3D head avatars, providing greater flexibility than mesh-based methods and more efficient rendering compared to NeRF-based approaches. Despite these advancements, the creation of controllable 3DGS-based head avatars remains time-intensive, often requiring tens of minutes to hours. To expedite this process, we here introduce the "Gaussian Deja-vu" framework, which first obtains a generalized model of the head avatar and then personalizes the result. The generalized model is trained on large 2D (synthetic and real) image datasets. This model provides a well-initialized 3D Gaussian head that is further refined using a monocular video to achieve the personalized head avatar. For personalizing, we propose learnable expression-aware rectification blendmaps to correct the initial 3D Gaussians, ensuring rapid convergence without the reliance on neural networks. Experiments demonstrate that the proposed method meets its objectives. It outperforms state-of-the-art 3D Gaussian head avatars in terms of photorealistic quality as well as reduces training time consumption to at least a quarter of the existing methods, producing the avatar in minutes.

Mohsen Gholami, Rabab Ward, Z. Jane Wang

This paper proposes an end-to-end framework for generating 3D human pose datasets using Neural Radiance Fields (NeRF). Public datasets generally have limited diversity in terms of human poses and camera viewpoints, largely due to the resource-intensive nature of collecting 3D human pose data. As a result, pose estimators trained on public datasets significantly underperform when applied to unseen out-of-distribution samples. Previous works proposed augmenting public datasets by generating 2D-3D pose pairs or rendering a large amount of random data. Such approaches either overlook image rendering or result in suboptimal datasets for pre-trained models. Here we propose PoseGen, which learns to generate a dataset (human 3D poses and images) with a feedback loss from a given pre-trained pose estimator. In contrast to prior art, our generated data is optimized to improve the robustness of the pre-trained model. The objective of PoseGen is to learn a distribution of data that maximizes the prediction error of a given pre-trained model. As the learned data distribution contains OOD samples of the pre-trained model, sampling data from such a distribution for further fine-tuning a pre-trained model improves the generalizability of the model. This is the first work that proposes NeRFs for 3D human data generation. NeRFs are data-driven and do not require 3D scans of humans. Therefore, using NeRF for data generation is a new direction for convenient user-specific data generation. Our extensive experiments show that the proposed PoseGen improves two baseline models (SPIN and HybrIK) on four datasets with an average 6% relative improvement.

Peizhi Yan, Rabab Ward, Qiang Tang et al.

3D Gaussian Splatting (3DGS) has enabled photorealistic and real-time rendering of 3D head avatars. Existing 3DGS-based avatars typically rely on tens of thousands of 3D Gaussian points (Gaussians), with the number of Gaussians fixed after training. However, many practical applications require adjustable levels of detail (LOD) to balance rendering efficiency and visual quality. In this work, we propose "ArchitectHead", the first framework for creating 3D Gaussian head avatars that support continuous control over LOD. Our key idea is to parameterize the Gaussians in a 2D UV feature space and propose a UV feature field composed of multi-level learnable feature maps to encode their latent features. A lightweight neural network-based decoder then transforms these latent features into 3D Gaussian attributes for rendering. ArchitectHead controls the number of Gaussians by dynamically resampling feature maps from the UV feature field at the desired resolutions. This method enables efficient and continuous control of LOD without retraining. Experimental results show that ArchitectHead achieves state-of-the-art (SOTA) quality in self and cross-identity reenactment tasks at the highest LOD, while maintaining near SOTA performance at lower LODs. At the lowest LOD, our method uses only 6.2\% of the Gaussians while the quality degrades moderately (L1 Loss +7.9\%, PSNR --0.97\%, SSIM --0.6\%, LPIPS Loss +24.1\%), and the rendering speed nearly doubles.

Mohsen Gholami, Bastian Wandt, Helge Rhodin et al.

This paper addresses the problem of cross-dataset generalization of 3D human pose estimation models. Testing a pre-trained 3D pose estimator on a new dataset results in a major performance drop. Previous methods have mainly addressed this problem by improving the diversity of the training data. We argue that diversity alone is not sufficient and that the characteristics of the training data need to be adapted to those of the new dataset such as camera viewpoint, position, human actions, and body size. To this end, we propose AdaptPose, an end-to-end framework that generates synthetic 3D human motions from a source dataset and uses them to fine-tune a 3D pose estimator. AdaptPose follows an adversarial training scheme. From a source 3D pose the generator generates a sequence of 3D poses and a camera orientation that is used to project the generated poses to a novel view. Without any 3D labels or camera information AdaptPose successfully learns to create synthetic 3D poses from the target dataset while only being trained on 2D poses. In experiments on the Human3.6M, MPI-INF-3DHP, 3DPW, and Ski-Pose datasets our method outperforms previous work in cross-dataset evaluations by 14% and previous semi-supervised learning methods that use partial 3D annotations by 16%.

Mohsen Gholami, Ahmad Rezaei, Helge Rhodin et al.

Estimating 3D human poses from video is a challenging problem. The lack of 3D human pose annotations is a major obstacle for supervised training and for generalization to unseen datasets. In this work, we address this problem by proposing a weakly-supervised training scheme that does not require 3D annotations or calibrated cameras. The proposed method relies on temporal information and triangulation. Using 2D poses from multiple views as the input, we first estimate the relative camera orientations and then generate 3D poses via triangulation. The triangulation is only applied to the views with high 2D human joint confidence. The generated 3D poses are then used to train a recurrent lifting network (RLN) that estimates 3D poses from 2D poses. We further apply a multi-view re-projection loss to the estimated 3D poses and enforce the 3D poses estimated from multi-views to be consistent. Therefore, our method relaxes the constraints in practice, only multi-view videos are required for training, and is thus convenient for in-the-wild settings. At inference, RLN merely requires single-view videos. The proposed method outperforms previous works on two challenging datasets, Human3.6M and MPI-INF-3DHP. Codes and pretrained models will be publicly available.

Dan Wang, Xinrui Cui, Xun Chen et al.

Deep CNN-based methods have so far achieved the state of the art results in multi-view 3D object reconstruction. Despite the considerable progress, the two core modules of these methods - multi-view feature extraction and fusion, are usually investigated separately, and the object relations in different views are rarely explored. In this paper, inspired by the recent great success in self-attention-based Transformer models, we reformulate the multi-view 3D reconstruction as a sequence-to-sequence prediction problem and propose a new framework named 3D Volume Transformer (VolT) for such a task. Unlike previous CNN-based methods using a separate design, we unify the feature extraction and view fusion in a single Transformer network. A natural advantage of our design lies in the exploration of view-to-view relationships using self-attention among multiple unordered inputs. On ShapeNet - a large-scale 3D reconstruction benchmark dataset, our method achieves a new state-of-the-art accuracy in multi-view reconstruction with fewer parameters ($70\%$ less) than other CNN-based methods. Experimental results also suggest the strong scaling capability of our method. Our code will be made publicly available.

Mazen Abdelfattah, Kaiwen Yuan, Z. Jane Wang et al.

Most autonomous vehicles (AVs) rely on LiDAR and RGB camera sensors for perception. Using these point cloud and image data, perception models based on deep neural nets (DNNs) have achieved state-of-the-art performance in 3D detection. The vulnerability of DNNs to adversarial attacks has been heavily investigated in the RGB image domain and more recently in the point cloud domain, but rarely in both domains simultaneously. Multi-modal perception systems used in AVs can be divided into two broad types: cascaded models which use each modality independently, and fusion models which learn from different modalities simultaneously. We propose a universal and physically realizable adversarial attack for each type, and study and contrast their respective vulnerabilities to attacks. We place a single adversarial object with specific shape and texture on top of a car with the objective of making this car evade detection. Evaluating on the popular KITTI benchmark, our adversarial object made the host vehicle escape detection by each model type more than 50% of the time. The dense RGB input contributed more to the success of the adversarial attacks on both cascaded and fusion models.

Davood Karimi, Rabab K. Ward

As the medical usage of computed tomography (CT) continues to grow, the radiation dose should remain at a low level to reduce the health risks. Therefore, there is an increasing need for algorithms that can reconstruct high-quality images from low-dose scans. In this regard, most of the recent studies have focused on iterative reconstruction algorithms, and little attention has been paid to restoration of the projection measurements, i.e., the sinogram. In this paper, we propose a novel sinogram interpolation algorithm. The proposed algorithm exploits the self-similarity and smoothness of the sinogram. Sinogram self-similarity is modeled in terms of the similarity of small blocks extracted from stacked projections. The smoothness is modeled via second-order total variation. Experiments with simulated and real CT data show that sinogram interpolation with the proposed algorithm leads to a substantial improvement in the quality of the reconstructed image, especially on low-dose scans. The proposed method can result in a significant reduction in the number of projection measurements. This will reduce the radiation dose and also the amount of data that need to be stored or transmitted, if the reconstruction is to be performed in a remote site.

Mazen Abdelfattah, Kaiwen Yuan, Z. Jane Wang et al.

We propose a universal and physically realizable adversarial attack on a cascaded multi-modal deep learning network (DNN), in the context of self-driving cars. DNNs have achieved high performance in 3D object detection, but they are known to be vulnerable to adversarial attacks. These attacks have been heavily investigated in the RGB image domain and more recently in the point cloud domain, but rarely in both domains simultaneously - a gap to be filled in this paper. We use a single 3D mesh and differentiable rendering to explore how perturbing the mesh's geometry and texture can reduce the robustness of DNNs to adversarial attacks. We attack a prominent cascaded multi-modal DNN, the Frustum-Pointnet model. Using the popular KITTI benchmark, we showed that the proposed universal multi-modal attack was successful in reducing the model's ability to detect a car by nearly 73%. This work can aid in the understanding of what the cascaded RGB-point cloud DNN learns and its vulnerability to adversarial attacks.

Yongwei Wang, Mingquan Feng, Rabab Ward et al.

Deep neural networks are vulnerable to adversarial attacks. White-box adversarial attacks can fool neural networks with small adversarial perturbations, especially for large size images. However, keeping successful adversarial perturbations imperceptible is especially challenging for transfer-based black-box adversarial attacks. Often such adversarial examples can be easily spotted due to their unpleasantly poor visual qualities, which compromises the threat of adversarial attacks in practice. In this study, to improve the image quality of black-box adversarial examples perceptually, we propose structure-aware adversarial attacks by generating adversarial images based on psychological perceptual models. Specifically, we allow higher perturbations on perceptually insignificant regions, while assigning lower or no perturbation on visually sensitive regions. In addition to the proposed spatial-constrained adversarial perturbations, we also propose a novel structure-aware frequency adversarial attack method in the discrete cosine transform (DCT) domain. Since the proposed attacks are independent of the gradient estimation, they can be directly incorporated with existing gradient-based attacks. Experimental results show that, with the comparable attack success rate (ASR), the proposed methods can produce adversarial examples with considerably improved visual quality for free. With the comparable perceptual quality, the proposed approaches achieve higher attack success rates: particularly for the frequency structure-aware attacks, the average ASR improves more than 10% over the baseline attacks.

Yongwei Wang, Xin Ding, Li Ding et al.

Recently, generative adversarial networks (GANs) can generate photo-realistic fake facial images which are perceptually indistinguishable from real face photos, promoting research on fake face detection. Though fake face forensics can achieve high detection accuracy, their anti-forensic counterparts are less investigated. Here we explore more \textit{imperceptible} and \textit{transferable} anti-forensics for fake face imagery detection based on adversarial attacks. Since facial and background regions are often smooth, even small perturbation could cause noticeable perceptual impairment in fake face images. Therefore it makes existing adversarial attacks ineffective as an anti-forensic method. Our perturbation analysis reveals the intuitive reason of the perceptual degradation issue when directly applying existing attacks. We then propose a novel adversarial attack method, better suitable for image anti-forensics, in the transformed color domain by considering visual perception. Simple yet effective, the proposed method can fool both deep learning and non-deep learning based forensic detectors, achieving higher attack success rate and significantly improved visual quality. Specially, when adversaries consider imperceptibility as a constraint, the proposed anti-forensic method can improve the average attack success rate by around 30\% on fake face images over two baseline attacks. \textit{More imperceptible} and \textit{more transferable}, the proposed method raises new security concerns to fake face imagery detection. We have released our code for public use, and hopefully the proposed method can be further explored in related forensic applications as an anti-forensic benchmark.

Anupriya Gogna, Angshul Majumdar, Rabab Ward

In this work we propose an autoencoder based framework for simultaneous reconstruction and classification of biomedical signals. Previously these two tasks, reconstruction and classification were treated as separate problems. This is the first work to propose a combined framework to address the issue in a holistic fashion. Reconstruction techniques for biomedical signals for tele-monitoring are largely based on compressed sensing (CS) based method, these are designed techniques where the reconstruction formulation is based on some assumption regarding the signal. In this work, we propose a new paradigm for reconstruction we learn to reconstruct. An autoencoder can be trained for the same. But since the final goal is to analyze classify the signal we learn a linear classification map inside the autoencoder. The ensuing optimization problem is solved using the Split Bregman technique. Experiments have been carried out on reconstruction and classification of ECG arrhythmia classification and EEG seizure classification signals. Our proposed tool is capable of operating in a semi-supervised fashion. We show that our proposed method is better and more than an order magnitude faster in reconstruction than CS based methods; it is capable of real-time operation. Our method is also better than recently proposed classification methods. Significance: This is the first work offering an alternative to CS based reconstruction. It also shows that representation learning can yield better results than hand-crafted features for signal analysis.

Ramy Hussein, Mohamed Osama Ahmed, Rabab Ward et al.

Objective: The aim of this study is to develop an efficient and reliable epileptic seizure prediction system using intracranial EEG (iEEG) data, especially for people with drug-resistant epilepsy. The prediction procedure should yield accurate results in a fast enough fashion to alert patients of impending seizures. Methods: We quantitatively analyze the human iEEG data to obtain insights into how the human brain behaves before and between epileptic seizures. We then introduce an efficient pre-processing method for reducing the data size and converting the time-series iEEG data into an image-like format that can be used as inputs to convolutional neural networks (CNNs). Further, we propose a seizure prediction algorithm that uses cooperative multi-scale CNNs for automatic feature learning of iEEG data. Results: 1) iEEG channels contain complementary information and excluding individual channels is not advisable to retain the spatial information needed for accurate prediction of epileptic seizures. 2) The traditional PCA is not a reliable method for iEEG data reduction in seizure prediction. 3) Hand-crafted iEEG features may not be suitable for reliable seizure prediction performance as the iEEG data varies between patients and over time for the same patient. 4) Seizure prediction results show that our algorithm outperforms existing methods by achieving an average sensitivity of 87.85% and AUC score of 0.84. Conclusion: Understanding how the human brain behaves before seizure attacks and far from them facilitates better designs of epileptic seizure predictors. Significance: Accurate seizure prediction algorithms can warn patients about the next seizure attack so they could avoid dangerous activities. Medications could then be administered to abort the impending seizure and minimize the risk of injury.

Jianzhe Lin, Qi Wang, Rabab Ward et al.

Previous transfer learning methods based on deep network assume the knowledge should be transferred between the same hidden layers of the source domain and the target domains. This assumption doesn't always hold true, especially when the data from the two domains are heterogeneous with different resolutions. In such case, the most suitable numbers of layers for the source domain data and the target domain data would differ. As a result, the high level knowledge from the source domain would be transferred to the wrong layer of target domain. Based on this observation, "where to transfer" proposed in this paper should be a novel research frontier. We propose a new mathematic model named DT-LET to solve this heterogeneous transfer learning problem. In order to select the best matching of layers to transfer knowledge, we define specific loss function to estimate the corresponding relationship between high-level features of data in the source domain and the target domain. To verify this proposed cross-layer model, experiments for two cross-domain recognition/classification tasks are conducted, and the achieved superior results demonstrate the necessity of layer correspondence searching.

Fahime Sheikhzadeh, Martial Guillaud, Rabab K. Ward

This paper addresses the problem of quantifying biomarkers in multi-stained tissues, based on color and spatial information. A deep learning based method that can automatically localize and quantify the cells expressing biomarker(s) in a whole slide image is proposed. The deep learning network is a fully convolutional network (FCN) whose input is the true RGB color image of a tissue and output is a map of the different biomarkers. The FCN relies on a convolutional neural network (CNN) that classifies each cell separately according to the biomarker it expresses. In this study, images of immunohistochemistry (IHC) stained slides were collected and used. More than 4,500 RGB images of cells were manually labeled based on the expressing biomarkers. The labeled cell images were used to train the CNN (obtaining an accuracy of 92% in a test set). The trained CNN is then extended to an FCN that generates a map of all biomarkers in the whole slide image acquired by the scanner (instead of classifying every cell image). To evaluate our method, we manually labeled all nuclei expressing different biomarkers in two whole slide images and used theses as the ground truth. Our proposed method for immunohistochemical analysis compares well with the manual labeling by humans (average F-score of 0.96).

Hamid Palangi, Rabab Ward, Li Deng

Various studies that address the compressed sensing problem with Multiple Measurement Vectors (MMVs) have been recently carried. These studies assume the vectors of the different channels to be jointly sparse. In this paper, we relax this condition. Instead we assume that these sparse vectors depend on each other but that this dependency is unknown. We capture this dependency by computing the conditional probability of each entry in each vector being non-zero, given the "residuals" of all previous vectors. To estimate these probabilities, we propose the use of the Long Short-Term Memory (LSTM)[1], a data driven model for sequence modelling that is deep in time. To calculate the model parameters, we minimize a cross entropy cost function. To reconstruct the sparse vectors at the decoder, we propose a greedy solver that uses the above model to estimate the conditional probabilities. By performing extensive experiments on two real world datasets, we show that the proposed method significantly outperforms the general MMV solver (the Simultaneous Orthogonal Matching Pursuit (SOMP)) and a number of the model-based Bayesian methods. The proposed method does not add any complexity to the general compressive sensing encoder. The trained model is used just at the decoder. As the proposed method is a data driven method, it is only applicable when training data is available. In many applications however, training data is indeed available, e.g. in recorded images and videos.

H. Palangi, L. Deng, Y. Shen et al.

In this paper we address the following problem in web document and information retrieval (IR): How can we use long-term context information to gain better IR performance? Unlike common IR methods that use bag of words representation for queries and documents, we treat them as a sequence of words and use long short term memory (LSTM) to capture contextual dependencies. To the best of our knowledge, this is the first time that LSTM is applied to information retrieval tasks. Unlike training traditional LSTMs, the training strategy is different due to the special nature of information retrieval problem. Experimental evaluation on an IR task derived from the Bing web search demonstrates the ability of the proposed method in addressing both lexical mismatch and long-term context modelling issues, thereby, significantly outperforming existing state of the art methods for web document retrieval task.

Hamid Palangi, Li Deng, Rabab K Ward

Echo State Networks (ESNs) are a special type of the temporally deep network model, the Recurrent Neural Network (RNN), where the recurrent matrix is carefully designed and both the recurrent and input matrices are fixed. An ESN uses the linearity of the activation function of the output units to simplify the learning of the output matrix. In this paper, we devise a special technique that take advantage of this linearity in the output units of an ESN, to learn the input and recurrent matrices. This has not been done in earlier ESNs due to their well known difficulty in learning those matrices. Compared to the technique of BackPropagation Through Time (BPTT) in learning general RNNs, our proposed method exploits linearity of activation function in the output units to formulate the relationships amongst the various matrices in an RNN. These relationships results in the gradient of the cost function having an analytical form and being more accurate. This would enable us to compute the gradients instead of obtaining them by recursion as in BPTT. Experimental results on phone state classification show that learning one or both the input and recurrent matrices in an ESN yields superior results compared to traditional ESNs that do not learn these matrices, especially when longer time steps are used.

Tanaya Guha, Ehsan Nezhadarya, Rabab K Ward

A successful approach to image quality assessment involves comparing the structural information between a distorted and its reference image. However, extracting structural information that is perceptually important to our visual system is a challenging task. This paper addresses this issue by employing a sparse representation-based approach and proposes a new metric called the \emph{sparse representation-based quality} (SPARQ) \emph{index}. The proposed method learns the inherent structures of the reference image as a set of basis vectors, such that any structure in the image can be represented by a linear combination of only a few of those basis vectors. This sparse strategy is employed because it is known to generate basis vectors that are qualitatively similar to the receptive field of the simple cells present in the mammalian primary visual cortex. The visual quality of the distorted image is estimated by comparing the structures of the reference and the distorted images in terms of the learnt basis vectors resembling cortical cells. Our approach is evaluated on six publicly available subject-rated image quality assessment datasets. The proposed SPARQ index consistently exhibits high correlation with the subjective ratings on all datasets and performs better or at par with the state-of-the-art.

Mani Malekesmaeili, Rabab K. Ward

This study proposes an audio copy detection system that is robust to various attacks. These include the severe pitch shift and tempo change attacks which existing systems fail to detect. First, we propose a novel two dimensional representation for audio signals called the time-chroma image. This image is based on a modification of the concept of chroma in the music literature and is shown to achieve better performance in song identification. Then, we propose a novel fingerprinting algorithm that extracts local fingerprints from the time-chroma image. The proposed local fingerprinting algorithm is invariant to time/frequency scale changes in audio signals. It also outperforms existing methods like SIFT by a great extent. Finally, we introduce a song identification algorithm that uses the proposed fingerprints. The resulting copy detection system is shown to significantly outperform existing methods. Besides being able to detect whether a song (or a part of it) has been copied, the proposed system can accurately estimate the amount of pitch shift and/or tempo change that might have been applied to a song.

Tanaya Guha, Rabab K. Ward

A new line of research uses compression methods to measure the similarity between signals. Two signals are considered similar if one can be compressed significantly when the information of the other is known. The existing compression-based similarity methods, although successful in the discrete one dimensional domain, do not work well in the context of images. This paper proposes a sparse representation-based approach to encode the information content of an image using information from the other image, and uses the compactness (sparsity) of the representation as a measure of its compressibility (how much can the image be compressed) with respect to the other image. The more sparse the representation of an image, the better it can be compressed and the more it is similar to the other image. The efficacy of the proposed measure is demonstrated through the high accuracies achieved in image clustering, retrieval and classification.