Analysis of cardiac ultrasound images is commonly performed in routine clinical practice for quantification of cardiac function. Its increasing automation frequently employs deep learning networks that are trained to predict disease or detect image features. However, such models are extremely data-hungry and training requires labelling of many thousands of images by experienced clinicians. Here we propose the use of contrastive learning to mitigate the labelling bottleneck. We train view classification models for imbalanced cardiac ultrasound datasets and show improved performance for views/classes for which minimal labelled data is available. Compared to a naive baseline model, we achieve an improvement in F1 score of up to 26% in those views while maintaining state-of-the-art performance for the views with sufficiently many labelled training observations.
We tackle the low-efficiency flaw of vision transformer caused by the high computational/space complexity in Multi-Head Self-Attention (MHSA). To this end, we propose the Hierarchical MHSA (H-MHSA), whose representation is computed in a hierarchical manner. Specifically, our H-MHSA first learns feature relationships within small grids by viewing image patches as tokens. Then, small grids are merged into larger ones, within which feature relationship is learned by viewing each small grid at the preceding step as a token. This process is iterated to gradually reduce the number of tokens. The H-MHSA module is readily pluggable into any CNN architectures and amenable to training via backpropagation. We call this new backbone TransCNN, and it essentially inherits the advantages of both transformer and CNN. Experiments demonstrate that TransCNN achieves state-of-the-art accuracy for image recognition. Code and pretrained models are available at https://github.com/yun-liu/TransCNN. This technical report will keep updating by adding more experiments.
Multilingual (or cross-lingual) embeddings represent several languages in a unique vector space. Using a common embedding space enables for a shared semantic between words from different languages. In this paper, we propose to embed images and texts into a unique distributional vector space, enabling to search images by using text queries expressing information needs related to the (visual) content of images, as well as using image similarity. Our framework forces the representation of an image to be similar to the representation of the text that describes it. Moreover, by using multilingual embeddings we ensure that words from two different languages have close descriptors and thus are attached to similar images. We provide experimental evidence of the efficiency of our approach by experimenting it on two datasets: Common Objects in COntext (COCO) [19] and Multi30K [7].
We introduce a novel fine-grained dataset and benchmark, the Danish Fungi 2020 (DF20). The dataset, constructed from observations submitted to the Danish Fungal Atlas, is unique in its taxonomy-accurate class labels, small number of errors, highly unbalanced long-tailed class distribution, rich observation metadata, and well-defined class hierarchy. DF20 has zero overlap with ImageNet, allowing unbiased comparison of models fine-tuned from publicly available ImageNet checkpoints. The proposed evaluation protocol enables testing the ability to improve classification using metadata -- e.g. precise geographic location, habitat, and substrate, facilitates classifier calibration testing, and finally allows to study the impact of the device settings on the classification performance. Experiments using Convolutional Neural Networks (CNN) and the recent Vision Transformers (ViT) show that DF20 presents a challenging task. Interestingly, ViT achieves results superior to CNN baselines with 81.25% accuracy, reducing the CNN error by 13%. A baseline procedure for including metadata into the decision process improves the classification accuracy by more than 3.5 percentage points, reducing the error rate by 20%. The source code for all methods and experiments is available at https://sites.google.com/view/danish-fungi-dataset.
As an important task in multimodal context understanding, Text-VQA (Visual Question Answering) aims at question answering through reading text information in images. It differentiates from the original VQA task as Text-VQA requires large amounts of scene-text relationship understanding, in addition to the cross-modal grounding capability. In this paper, we propose Localize, Group, and Select (LOGOS), a novel model which attempts to tackle this problem from multiple aspects. LOGOS leverages two grounding tasks to better localize the key information of the image, utilizes scene text clustering to group individual OCR tokens, and learns to select the best answer from different sources of OCR (Optical Character Recognition) texts. Experiments show that LOGOS outperforms previous state-of-the-art methods on two Text-VQA benchmarks without using additional OCR annotation data. Ablation studies and analysis demonstrate the capability of LOGOS to bridge different modalities and better understand scene text.
Classic and deep learning-based generalized canonical correlation analysis (GCCA) algorithms seek low-dimensional common representations of data entities from multiple ``views'' (e.g., audio and image) using linear transformations and neural networks, respectively. When the views are acquired and stored at different locations, organizations and edge devices, computing GCCA in a distributed, parallel and efficient manner is well-motivated. However, existing distributed GCCA algorithms may incur prohitively high communication overhead. This work puts forth a communication-efficient distributed framework for both linear and deep GCCA under the maximum variance (MAX-VAR) paradigm. The overhead issue is addressed by aggressively compressing (via quantization) the exchanging information between the distributed computing agents and a central controller. Compared to the unquantized version, the proposed algorithm consistently reduces the communication overhead by about $90\%$ with virtually no loss in accuracy and convergence speed. Rigorous convergence analyses are also presented -- which is a nontrivial effort since no existing generic result from quantized distributed optimization covers the special problem structure of GCCA. Our result shows that the proposed algorithms for both linear and deep GCCA converge to critical points in a sublinear rate, even under heavy quantization and stochastic approximations. In addition, it is shown that in the linear MAX-VAR case, the quantized algorithm approaches a {\it global optimum} in a {\it geometric} rate -- if the computing agents' updates meet a certain accuracy level. Synthetic and real data experiments are used to showcase the effectiveness of the proposed approach.
3D human body reconstruction from monocular images is an interesting and ill-posed problem in computer vision with wider applications in multiple domains. In this paper, we propose SHARP, a novel end-to-end trainable network that accurately recovers the detailed geometry and appearance of 3D people in loose clothing from a monocular image. We propose a sparse and efficient fusion of a parametric body prior with a non-parametric peeled depth map representation of clothed models. The parametric body prior constraints our model in two ways: first, the network retains geometrically consistent body parts that are not occluded by clothing, and second, it provides a body shape context that improves prediction of the peeled depth maps. This enables SHARP to recover fine-grained 3D geometrical details with just L1 losses on the 2D maps, given an input image. We evaluate SHARP on publicly available Cloth3D and THuman datasets and report superior performance to state-of-the-art approaches.
For the majority of the learning-based segmentation methods, a large quantity of high-quality training data is required. In this paper, we present a novel learning-based segmentation model that could be trained semi- or un- supervised. Specifically, in the unsupervised setting, we parameterize the Active contour without edges (ACWE) framework via a convolutional neural network (ConvNet), and optimize the parameters of the ConvNet using a self-supervised method. In another setting (semi-supervised), the auxiliary segmentation ground truth is used during training. We show that the method provides fast and high-quality bone segmentation in the context of single-photon emission computed tomography (SPECT) image.
Recently, deep learning based single image super-resolution(SR) approaches have achieved great development. The state-of-the-art SR methods usually adopt a feed-forward pipeline to establish a non-linear mapping between low-res(LR) and high-res(HR) images. However, due to treating all image regions equally without considering the difficulty diversity, these approaches meet an upper bound for optimization. To address this issue, we propose a novel SR approach that discriminately processes each image region within an image by its difficulty. Specifically, we propose a dual-way SR network that one way is trained to focus on easy image regions and another is trained to handle hard image regions. To identify whether a region is easy or hard, we propose a novel image difficulty recognition network based on PSNR prior. Our SR approach that uses the region mask to adaptively enforce the dual-way SR network yields superior results. Extensive experiments on several standard benchmarks (e.g., Set5, Set14, BSD100, and Urban100) show that our approach achieves state-of-the-art performance.
The vision community is witnessing a modeling shift from CNNs to Transformers, where pure Transformer architectures have attained top accuracy on the major video recognition benchmarks. These video models are all built on Transformer layers that globally connect patches across the spatial and temporal dimensions. In this paper, we instead advocate an inductive bias of locality in video Transformers, which leads to a better speed-accuracy trade-off compared to previous approaches which compute self-attention globally even with spatial-temporal factorization. The locality of the proposed video architecture is realized by adapting the Swin Transformer designed for the image domain, while continuing to leverage the power of pre-trained image models. Our approach achieves state-of-the-art accuracy on a broad range of video recognition benchmarks, including on action recognition (84.9 top-1 accuracy on Kinetics-400 and 86.1 top-1 accuracy on Kinetics-600 with ~20x less pre-training data and ~3x smaller model size) and temporal modeling (69.6 top-1 accuracy on Something-Something v2). The code and models will be made publicly available at https://github.com/SwinTransformer/Video-Swin-Transformer.