Dense panoptic prediction is a key ingredient in many existing applications such as autonomous driving, automated warehouses or agri-robotics. However, most of these applications leverage the recovered dense semantics as an input to visual closed-loop control. Hence, practical deployments require real-time inference over large input resolutions on embedded hardware. These requirements call for computationally efficient approaches which deliver high accuracy with limited computational resources. We propose to achieve this goal by trading-off backbone capacity for multi-scale feature extraction. In comparison with contemporaneous approaches to panoptic segmentation, the main novelties of our method are scale-equivariant feature extraction and cross-scale upsampling through pyramidal fusion. Our best model achieves 55.9% PQ on Cityscapes val at 60 FPS on full resolution 2MPx images and RTX3090 with FP16 Tensor RT optimization.
Standard machine learning is unable to accommodate inputs which do not belong to the training distribution. The resulting models often give rise to confident incorrect predictions which may lead to devastating consequences. This problem is especially demanding in the context of dense prediction since input images may be partially anomalous. Previous work has addressed dense anomaly detection by discriminative training on mixed-content images. We extend this approach with synthetic negative patches which simultaneously achieve high inlier likelihood and uniform discriminative prediction. We generate synthetic negatives with normalizing flows due to their outstanding distribution coverage and capability to generate samples at different resolutions. We also propose to detect anomalies according to a principled information-theoretic criterion which can be consistently applied through training and inference. The resulting models set the new state of the art on standard benchmarks and datasets in spite of minimal computational overhead and refraining from auxiliary negative data.
Deep supervised models have an unprecedented capacity to absorb large quantities of training data. Hence, training on many datasets becomes a method of choice towards graceful degradation in unusual scenes. Unfortunately, different datasets often use incompatible labels. For instance, the Cityscapes road class subsumes all driving surfaces, while Vistas defines separate classes for road markings, manholes etc. We address this challenge by proposing a principled method for seamless learning on datasets with overlapping classes based on partial labels and probabilistic loss. Our method achieves competitive within-dataset and cross-dataset generalization, as well as ability to learn visual concepts which are not separately labeled in any of the training datasets. Experiments reveal competitive or state-of-the-art performance on two multi-domain dataset collections and on the WildDash 2 benchmark.
Semi-supervised learning is especially interesting in the dense prediction context due to high cost of pixel-level ground truth. Unfortunately, most such approaches are evaluated on outdated architectures which hamper research due to very slow training and high requirements on GPU RAM. We address this concern by presenting a simple and effective baseline which works very well both on standard and efficient architectures. Our baseline is based on one-way consistency and non-linear geometric and photometric perturbations. We show advantage of perturbing only the student branch and present a plausible explanation of such behaviour. Experiments on Cityscapes and CIFAR-10 demonstrate competitive performance with respect to prior work.
Normalizing flows are bijective mappings between inputs and latent representations with a fully factorized distribution. They are very attractive due to exact likelihood evaluation and efficient sampling. However, their effective capacity is often insufficient since the bijectivity constraint limits the model width. We address this issue by incrementally padding intermediate representations with noise. We precondition the noise in accordance with previous invertible units, which we describe as cross-unit coupling. Our invertible glow-like modules express intra-unit affine coupling as a fusion of a densely connected block and Nystr\"om self-attention. We refer to our architecture as DenseFlow since both cross-unit and intra-unit couplings rely on dense connectivity. Experiments show significant improvements due to the proposed contributions, and reveal state-of-the-art density estimation among all generative models under moderate computing budgets.
We present a novel dense semantic forecasting approach which is applicable to a variety of architectures and tasks. The approach consists of two modules. Feature-to-motion (F2M) module forecasts a dense deformation field which warps past features into their future positions. Feature-to-feature (F2F) module regresses the future features directly and is therefore able to account for emergent scenery. The compound F2MF approach decouples effects of motion from the effects of novelty in a task-agnostic manner. We aim to apply F2MF forecasting to the most subsampled and the most abstract representation of a desired single-frame model. Our implementations take advantage of deformable convolutions and pairwise correlation coefficients across neighbouring time instants. We perform experiments on three dense prediction tasks: semantic segmentation, instance-level segmentation, and panoptic segmentation. The results reveal state-of-the-art forecasting accuracy across all three modalities on the Cityscapes dataset.
Deep convolutional models often produce inadequate predictions for inputs foreign to the training distribution. Consequently, the problem of detecting outlier images has recently been receiving a lot of attention. Unlike most previous work, we address this problem in the dense prediction context in order to be able to locate outlier objects in front of in-distribution background. Our approach is based on two reasonable assumptions. First, we assume that the inlier dataset is related to some narrow application field (e.g.~road driving). Second, we assume that there exists a general-purpose dataset which is much more diverse than the inlier dataset (e.g.~ImageNet-1k). We consider pixels from the general-purpose dataset as noisy negative training samples since most (but not all) of them are outliers. We encourage the model to recognize borders between known and unknown by pasting jittered negative patches over inlier training images. Our experiments target two dense open-set recognition benchmarks (WildDash 1 and Fishyscapes) and one dense open-set recognition dataset (StreetHazard). Extensive performance evaluation indicates competitive potential of the proposed approach.
Today's deep models are often unable to detect inputs which do not belong to the training distribution. This gives rise to confident incorrect predictions which could lead to devastating consequences in many important application fields such as healthcare and autonomous driving. Interestingly, both discriminative and generative models appear to be equally affected. Consequently, this vulnerability represents an important research challenge. We consider an outlier detection approach based on discriminative training with jointly learned synthetic outliers. We obtain the synthetic outliers by sampling an RNVP model which is jointly trained to generate datapoints at the border of the training distribution. We show that this approach can be adapted for simultaneous semantic segmentation and dense outlier detection. We present image classification experiments on CIFAR-10, as well as semantic segmentation experiments on three existing datasets (StreetHazards, WD-Pascal, Fishyscapes Lost & Found), and one contributed dataset. Our models perform competitively with respect to the state of the art despite producing predictions with only one forward pass.
This paper considers semantic forecasting in road-driving scenes. Most existing approaches address this problem as deterministic regression of future features or future predictions given observed frames. However, such approaches ignore the fact that future can not always be guessed with certainty. For example, when a car is about to turn around a corner, the road which is currently occluded by buildings may turn out to be either free to drive, or occupied by people, other vehicles or roadworks. When a deterministic model confronts such situation, its best guess is to forecast the most likely outcome. However, this is not acceptable since it defeats the purpose of forecasting to improve security. It also throws away valuable training data, since a deterministic model is unable to learn any deviation from the norm. We address this problem by providing more freedom to the model through allowing it to forecast different futures. We propose to formulate multimodal forecasting as sampling of a multimodal generative model conditioned on the observed frames. Experiments on the Cityscapes dataset reveal that our multimodal model outperforms its deterministic counterpart in short-term forecasting while performing slightly worse in the mid-term case.