Convolutional neural networks (CNNs) have recently achieved remarkable success in automatically identifying organs or lesions on 3D medical images. Meanwhile, vision transformer networks have exhibited exceptional performance in 2D image classification tasks. Compared with CNNs, transformer networks have an obvious advantage of extracting long-range features due to their self-attention algorithm. Therefore, in this paper we present a CNN-Transformer combined model called BiTr-Unet for brain tumor segmentation on multi-modal MRI scans. The proposed BiTr-Unet achieves good performance on the BraTS 2021 validation dataset with mean Dice score 0.9076, 0.8392 and 0.8231, and mean Hausdorff distance 4.5322, 13.4592 and 14.9963 for the whole tumor, tumor core, and enhancing tumor, respectively.
We investigate a general matrix factorization for deviance-based losses, extending the ubiquitous singular value decomposition beyond squared error loss. While similar approaches have been explored before, here we propose an efficient algorithm that is flexible enough to allow for structural zeros and entry weights. Moreover, we provide theoretical support for these decompositions by (i) showing strong consistency under a generalized linear model setup, (ii) checking the adequacy of a chosen exponential family via a generalized Hosmer-Lemeshow test, and (iii) determining the rank of the decomposition via a maximum eigenvalue gap method. To further support our findings, we conduct simulation studies to assess robustness to decomposition assumptions and extensive case studies using benchmark datasets from image face recognition, natural language processing, network analysis, and biomedical studies. Our theoretical and empirical results indicate that the proposed decomposition is more flexible, general, and can provide improved performance when compared to traditional methods.
We present a single-shot, bottom-up approach for whole image parsing. Whole image parsing, also known as Panoptic Segmentation, generalizes the tasks of semantic segmentation for 'stuff' classes and instance segmentation for 'thing' classes, assigning both semantic and instance labels to every pixel in an image. Recent approaches to whole image parsing typically employ separate standalone modules for the constituent semantic and instance segmentation tasks and require multiple passes of inference. Instead, the proposed DeeperLab image parser performs whole image parsing with a significantly simpler, fully convolutional approach that jointly addresses the semantic and instance segmentation tasks in a single-shot manner, resulting in a streamlined system that better lends itself to fast processing. For quantitative evaluation, we use both the instance-based Panoptic Quality (PQ) metric and the proposed region-based Parsing Covering (PC) metric, which better captures the image parsing quality on 'stuff' classes and larger object instances. We report experimental results on the challenging Mapillary Vistas dataset, in which our single model achieves 31.95% (val) / 31.6% PQ (test) and 55.26% PC (val) with 3 frames per second (fps) on GPU or near real-time speed (22.6 fps on GPU) with reduced accuracy.
We propose a general method to train a single convolutional neural network which is capable of switching image resolutions at inference. Thus the running speed can be selected to meet various computational resource limits. Networks trained with the proposed method are named Resolution Switchable Networks (RS-Nets). The basic training framework shares network parameters for handling images which differ in resolution, yet keeps separate batch normalization layers. Though it is parameter-efficient in design, it leads to inconsistent accuracy variations at different resolutions, for which we provide a detailed analysis from the aspect of the train-test recognition discrepancy. A multi-resolution ensemble distillation is further designed, where a teacher is learnt on the fly as a weighted ensemble over resolutions. Thanks to the ensemble and knowledge distillation, RS-Nets enjoy accuracy improvements at a wide range of resolutions compared with individually trained models. Extensive experiments on the ImageNet dataset are provided, and we additionally consider quantization problems. Code and models are available at https://github.com/yikaiw/RS-Nets.
The medical image is characterized by the inter-class indistinction, high variability, and noise, where the recognition of pixels is challenging. Unlike previous self-attention based methods that capture context information from one level, we reformulate the self-attention mechanism from the view of the high-order graph and propose a novel method, namely Hierarchical Attention Network (HANet), to address the problem of medical image segmentation. Concretely, an HA module embedded in the HANet captures context information from neighbors of multiple levels, where these neighbors are extracted from the high-order graph. In the high-order graph, there will be an edge between two nodes only if the correlation between them is high enough, which naturally reduces the noisy attention information caused by the inter-class indistinction. The proposed HA module is robust to the variance of input and can be flexibly inserted into the existing convolution neural networks. We conduct experiments on three medical image segmentation tasks including optic disc/cup segmentation, blood vessel segmentation, and lung segmentation. Extensive results show our method is more effective and robust than the existing state-of-the-art methods.
Infrared and visible image fusion expects to obtain images that highlight thermal radiation information from infrared images and texture details from visible images. In this paper, an interpretable deep network fusion model is proposed. Initially, two optimization models are established to accomplish two-scale decomposition, separating low-frequency base information and high-frequency detail information from source images. The algorithm unrolling that each iteration process is mapped to a convolutional neural network layer to transfer the optimization steps into the trainable neural networks, is implemented to solve the optimization models. In the test phase, the two decomposition feature maps of base and detail are merged respectively by the fusion layer, and then the decoder outputs the fusion image. Qualitative and quantitative comparisons demonstrate the superiority of our model, which is interpretable and can robustly generate fusion images containing highlight targets and legible details, exceeding the state-of-the-art methods.
Recently, a series of algorithms have been explored for GAN compression, which aims to reduce tremendous computational overhead and memory usages when deploying GANs on resource-constrained edge devices. However, most of the existing GAN compression work only focuses on how to compress the generator, while fails to take the discriminator into account. In this work, we revisit the role of discriminator in GAN compression and design a novel generator-discriminator cooperative compression scheme for GAN compression, termed GCC. Within GCC, a selective activation discriminator automatically selects and activates convolutional channels according to a local capacity constraint and a global coordination constraint, which help maintain the Nash equilibrium with the lightweight generator during the adversarial training and avoid mode collapse. The original generator and discriminator are also optimized from scratch, to play as a teacher model to progressively refine the pruned generator and the selective activation discriminator. A novel online collaborative distillation scheme is designed to take full advantage of the intermediate feature of the teacher generator and discriminator to further boost the performance of the lightweight generator. Extensive experiments on various GAN-based generation tasks demonstrate the effectiveness and generalization of GCC. Among them, GCC contributes to reducing 80% computational costs while maintains comparable performance in image translation tasks. Our code and models are available at https://github.com/SJLeo/GCC.
Discovering a solution in a combinatorial space is prevalent in many real-world problems but it is also challenging due to diverse complex constraints and the vast number of possible combinations. To address such a problem, we introduce a novel formulation, combinatorial construction, which requires a building agent to assemble unit primitives (i.e., LEGO bricks) sequentially -- every connection between two bricks must follow a fixed rule, while no bricks mutually overlap. To construct a target object, we provide incomplete knowledge about the desired target (i.e., 2D images) instead of exact and explicit volumetric information to the agent. This problem requires a comprehensive understanding of partial information and long-term planning to append a brick sequentially, which leads us to employ reinforcement learning. The approach has to consider a variable-sized action space where a large number of invalid actions, which would cause overlap between bricks, exist. To resolve these issues, our model, dubbed Brick-by-Brick, adopts an action validity prediction network that efficiently filters invalid actions for an actor-critic network. We demonstrate that the proposed method successfully learns to construct an unseen object conditioned on a single image or multiple views of a target object.
Deep generative models have shown success in automatically synthesizing missing image regions using surrounding context. However, users cannot directly decide what content to synthesize with such approaches. We propose an end-to-end network for image inpainting that uses a different image to guide the synthesis of new content to fill the hole. A key challenge addressed by our approach is synthesizing new content in regions where the guidance image and the context of the original image are inconsistent. We conduct four studies that demonstrate our results yield more realistic image inpainting results over seven baselines.
Lane segmentation is a challenging issue in autonomous driving system designing because lane marks show weak textural consistency due to occlusion or extreme illumination but strong geometric continuity in traffic images, from which general convolution neural networks (CNNs) are not capable of learning semantic objects. To empower conventional CNNs in learning geometric clues of lanes, we propose a deep network named ContinuityLearner to better learn geometric prior within lane. Specifically, our proposed CNN-based paradigm involves a novel Context-encoding image feature learning network to generate class-dependent image feature maps and a new encoding layer to exploit the geometric continuity feature representation by fusing both spatial and visual information of lane together. The ContinuityLearner, performing on the geometric continuity feature of lanes, is trained to directly predict the lane in traffic scenarios with integrated and continuous instance semantic. The experimental results on the CULane dataset and the Tusimple benchmark demonstrate that our ContinuityLearner has superior performance over other state-of-the-art techniques in lane segmentation.