Recent advances in self-supervised representation learning have enabled more efficient and robust model performance without relying on extensive labeled data. However, most works are still focused on images, with few working on videos and even fewer on multi-view videos, where more powerful inductive biases can be leveraged for self-supervision. In this work, we propose a novel method for representation learning of multi-view videos, where we explicitly model the representation space to maintain Homography Equivariance (HomE). Our method learns an implicit mapping between different views, culminating in a representation space that maintains the homography relationship between neighboring views. We evaluate our HomE representation via action recognition and pedestrian intent prediction as downstream tasks. On action classification, our method obtains 96.4% 3-fold accuracy on the UCF101 dataset, better than most state-of-the-art self-supervised learning methods. Similarly, on the STIP dataset, we outperform the state-of-the-art by 6% for pedestrian intent prediction one second into the future while also obtaining an accuracy of 91.2% for pedestrian action (cross vs. not-cross) classification. Code is available at https://github.com/anirudhs123/HomE.
Although the preservation of shape continuity and physiological anatomy is a natural assumption in the segmentation of medical images, it is often neglected by deep learning methods that mostly aim for the statistical modeling of input data as pixels rather than interconnected structures. In biological structures, however, organs are not separate entities; for example, in reality, a severed vessel is an indication of an underlying problem, but traditional segmentation models are not designed to strictly enforce the continuity of anatomy, potentially leading to inaccurate medical diagnoses. To address this issue, we propose a graph-based approach that enforces the continuity and connectivity of anatomical topology in medical images. Our method encodes the continuity of shapes as a graph constraint, ensuring that the network's predictions maintain this continuity. We evaluate our method on two public benchmarks on retinal vessel segmentation, showing significant improvements in connectivity metrics compared to traditional methods while getting better or on-par performance on segmentation metrics.
Although purely transformer-based architectures showed promising performance in many computer vision tasks, many hybrid models consisting of CNN and transformer blocks are introduced to fit more specialized tasks. Nevertheless, despite the performance gain of both pure and hybrid transformer-based architectures compared to CNNs in medical imaging segmentation, their high training cost and complexity make it challenging to use them in real scenarios. In this work, we propose simple architectures based on purely convolutional layers, and show that by just taking advantage of the attention map visualizations obtained from a self-supervised pretrained vision transformer network (e.g., DINO) one can outperform complex transformer-based networks with much less computation costs. The proposed architecture is composed of two encoder branches with the original image as input in one branch and the attention map visualizations of the same image from multiple self-attention heads from a pre-trained DINO model (as multiple channels) in the other branch. The results of our experiments on two publicly available medical imaging datasets show that the proposed pipeline outperforms U-Net and the state-of-the-art medical image segmentation models.
The effectiveness of computerized cognitive training in slowing cognitive decline and brain aging in dementia is often limited by the engagement of participants in the training. Monitoring older users' real-time engagement in domains of attention, motivation, and affect is crucial to understanding the overall effectiveness of such training. In this paper, we propose to predict engagement, quantified via an established mental fatigue measure assessing users' perceived attention, motivation, and affect throughout computerized cognitive training sessions, in older adults with mild cognitive impairment (MCI), by monitoring their real-time video-recorded facial gestures in training sessions. To achieve the goal, we used computer vision, analyzing video frames every 5 seconds to optimize the balance between information retention and data size, and developed a novel Recurrent Video Transformer (RVT). Our RVT model, which combines a clip-wise transformer encoder module and a session-wise Recurrent Neural Network (RNN) classifier, achieved the highest balanced accuracy, F1 score, and precision compared to other state-of-the-art models for both detecting mental fatigue/disengagement cases (binary classification) and rating the level of mental fatigue (multi-class classification). By leveraging dynamic temporal information, the RVT model demonstrates the potential to accurately predict engagement among computerized cognitive training users, which lays the foundation for future work to modulate the level of engagement in computerized cognitive training interventions. The code will be released.
Training deep learning models on brain MRI is often plagued by small sample size, which can lead to biased training or overfitting. One potential solution is to synthetically generate realistic MRIs via generative models such as Generative Adversarial Network (GAN). However, existing GANs for synthesizing realistic brain MRIs largely rely on image-to-image conditioned transformations requiring extensive, well-curated pairs of MRI samples for training. On the other hand, unconditioned GAN models (i.e., those generating MRI from random noise) are unstable during training and tend to produce blurred images during inference. Here, we propose an efficient strategy that generates high fidelity 3D brain MRI via Diffusion Probabilistic Model (DPM). To this end, we train a conditional DPM with attention to generate an MRI sub-volume (a set of slices at arbitrary locations) conditioned on another subset of slices from the same MRI. By computing attention weights from slice indices and using a mask to encode the target and conditional slices, the model is able to learn the long-range dependency across distant slices with limited computational resources. After training, the model can progressively synthesize a new 3D brain MRI by generating the first subset of slices from random noise and conditionally generating subsequent slices. Based on 1262 t1-weighted MRIs from three neuroimaging studies, our experiments demonstrate that the proposed method can generate high quality 3D MRIs that share the same distribution as real MRIs and are more realistic than the ones produced by GAN-based models.
Automatic medical image segmentation is a crucial topic in the medical domain and successively a critical counterpart in the computer-aided diagnosis paradigm. U-Net is the most widespread image segmentation architecture due to its flexibility, optimized modular design, and success in all medical image modalities. Over the years, the U-Net model achieved tremendous attention from academic and industrial researchers. Several extensions of this network have been proposed to address the scale and complexity created by medical tasks. Addressing the deficiency of the naive U-Net model is the foremost step for vendors to utilize the proper U-Net variant model for their business. Having a compendium of different variants in one place makes it easier for builders to identify the relevant research. Also, for ML researchers it will help them understand the challenges of the biological tasks that challenge the model. To address this, we discuss the practical aspects of the U-Net model and suggest a taxonomy to categorize each network variant. Moreover, to measure the performance of these strategies in a clinical application, we propose fair evaluations of some unique and famous designs on well-known datasets. We provide a comprehensive implementation library with trained models for future research. In addition, for ease of future studies, we created an online list of U-Net papers with their possible official implementation. All information is gathered in https://github.com/NITR098/Awesome-U-Net repository.
The white-matter (micro-)structural architecture of the brain promotes synchrony among neuronal populations, giving rise to richly patterned functional connections. A fundamental problem for systems neuroscience is determining the best way to relate structural and functional networks quantified by diffusion tensor imaging and resting-state functional MRI. As one of the state-of-the-art approaches for network analysis, graph convolutional networks (GCN) have been separately used to analyze functional and structural networks, but have not been applied to explore inter-network relationships. In this work, we propose to couple the two networks of an individual by adding inter-network edges between corresponding brain regions, so that the joint structure-function graph can be directly analyzed by a single GCN. The weights of inter-network edges are learnable, reflecting non-uniform structure-function coupling strength across the brain. We apply our Joint-GCN to predict age and sex of 662 participants from the public dataset of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) based on their functional and micro-structural white-matter networks. Our results support that the proposed Joint-GCN outperforms existing multi-modal graph learning approaches for analyzing structural and functional networks.
Human pose forecasting is a challenging problem involving complex human body motion and posture dynamics. In cases that there are multiple people in the environment, one's motion may also be influenced by the motion and dynamic movements of others. Although there are several previous works targeting the problem of multi-person dynamic pose forecasting, they often model the entire pose sequence as time series (ignoring the underlying relationship between joints) or only output the future pose sequence of one person at a time. In this paper, we present a new method, called Social Motion Transformer (SoMoFormer), for multi-person 3D pose forecasting. Our transformer architecture uniquely models human motion input as a joint sequence rather than a time sequence, allowing us to perform attention over joints while predicting an entire future motion sequence for each joint in parallel. We show that with this problem reformulation, SoMoFormer naturally extends to multi-person scenes by using the joints of all people in a scene as input queries. Using learned embeddings to denote the type of joint, person identity, and global position, our model learns the relationships between joints and between people, attending more strongly to joints from the same or nearby people. SoMoFormer outperforms state-of-the-art methods for long-term motion prediction on the SoMoF benchmark as well as the CMU-Mocap and MuPoTS-3D datasets. Code will be made available after publication.
There has been an increasing interest in multi-task learning for video understanding in recent years. In this work, we propose a generalized notion of multi-task learning by incorporating both auxiliary tasks that the model should perform well on and adversarial tasks that the model should not perform well on. We employ Necessary Condition Analysis (NCA) as a data-driven approach for deciding what category these tasks should fall in. Our novel proposed framework, Adversarial Multi-Task Neural Networks (AMT), penalizes adversarial tasks, determined by NCA to be scene recognition in the Holistic Video Understanding (HVU) dataset, to improve action recognition. This upends the common assumption that the model should always be encouraged to do well on all tasks in multi-task learning. Simultaneously, AMT still retains all the benefits of multi-task learning as a generalization of existing methods and uses object recognition as an auxiliary task to aid action recognition. We introduce two challenging Scene-Invariant test splits of HVU, where the model is evaluated on action-scene co-occurrences not encountered in training. We show that our approach improves accuracy by ~3% and encourages the model to attend to action features instead of correlation-biasing scene features.
For the first time, we propose using a multiple instance learning based convolution-free transformer model, called Multiple Instance Neuroimage Transformer (MINiT), for the classification of T1weighted (T1w) MRIs. We first present several variants of transformer models adopted for neuroimages. These models extract non-overlapping 3D blocks from the input volume and perform multi-headed self-attention on a sequence of their linear projections. MINiT, on the other hand, treats each of the non-overlapping 3D blocks of the input MRI as its own instance, splitting it further into non-overlapping 3D patches, on which multi-headed self-attention is computed. As a proof-of-concept, we evaluate the efficacy of our model by training it to identify sex from T1w-MRIs of two public datasets: Adolescent Brain Cognitive Development (ABCD) and the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA). The learned attention maps highlight voxels contributing to identifying sex differences in brain morphometry. The code is available at https://github.com/singlaayush/MINIT.