Although instance-aware perception is a key prerequisite for many autonomous robotic applications, most of the methods only partially solve the problem by focusing solely on known object categories. However, for robots interacting in dynamic and cluttered environments, this is not realistic and severely limits the range of potential applications. Therefore, we propose a novel object instance segmentation approach that does not require any semantic or geometric information of the objects beforehand. In contrast to existing works, we do not explicitly use depth data as input, but rely on the insight that slight viewpoint changes, which for example are provided by stereo image pairs, are often sufficient to determine object boundaries and thus to segment objects. Focusing on the versatility of stereo sensors, we employ a transformer-based architecture that maps directly from the pair of input images to the object instances. This has the major advantage that instead of a noisy, and potentially incomplete depth map as an input, on which the segmentation is computed, we use the original image pair to infer the object instances and a dense depth map. In experiments in several different application domains, we show that our Instance Stereo Transformer (INSTR) algorithm outperforms current state-of-the-art methods that are based on depth maps. Training code and pretrained models will be made available.
How to estimate the quality of the network output is an important issue, and currently there is no effective solution in the field of human parsing. In order to solve this problem, this work proposes a statistical method based on the output probability map to calculate the pixel quality information, which is called pixel score. In addition, the Quality-Aware Module (QAM) is proposed to fuse the different quality information, the purpose of which is to estimate the quality of human parsing results. We combine QAM with a concise and effective network design to propose Quality-Aware Network (QANet) for human parsing. Benefiting from the superiority of QAM and QANet, we achieve the best performance on three multiple and one single human parsing benchmarks, including CIHP, MHP-v2, Pascal-Person-Part and LIP. Without increasing the training and inference time, QAM improves the AP$^\text{r}$ criterion by more than 10 points in the multiple human parsing task. QAM can be extended to other tasks with good quality estimation, e.g. instance segmentation. Specifically, QAM improves Mask R-CNN by ~1% mAP on COCO and LVISv1.0 datasets. Based on the proposed QAM and QANet, our overall system wins 1st place in CVPR2019 COCO DensePose Challenge, and 1st place in Track 1 & 2 of CVPR2020 LIP Challenge. Code and models are available at https://github.com/soeaver/QANet.
Self-supervised learning has been widely used to obtain transferrable representations from unlabeled images. Especially, recent contrastive learning methods have shown impressive performances on downstream image classification tasks. While these contrastive methods mainly focus on generating invariant global representations at the image-level under semantic-preserving transformations, they are prone to overlook spatial consistency of local representations and therefore have a limitation in pretraining for localization tasks such as object detection and instance segmentation. Moreover, aggressively cropped views used in existing contrastive methods can minimize representation distances between the semantically different regions of a single image. In this paper, we propose a spatially consistent representation learning algorithm (SCRL) for multi-object and location-specific tasks. In particular, we devise a novel self-supervised objective that tries to produce coherent spatial representations of a randomly cropped local region according to geometric translations and zooming operations. On various downstream localization tasks with benchmark datasets, the proposed SCRL shows significant performance improvements over the image-level supervised pretraining as well as the state-of-the-art self-supervised learning methods. Code is available at https://github.com/kakaobrain/scrl
Classic computer vision algorithms, instance segmentation, and semantic segmentation can not provide a holistic understanding of the surroundings for the visually impaired. In this paper, we utilize panoptic segmentation to assist the navigation of visually impaired people by offering both things and stuff awareness in the proximity of the visually impaired efficiently. To this end, we propose an efficient Attention module -- Lintention which can model long-range interactions in linear time using linear space. Based on Lintention, we then devise a novel panoptic segmentation model which we term Panoptic Lintention Net. Experiments on the COCO dataset indicate that the Panoptic Lintention Net raises the Panoptic Quality (PQ) from 39.39 to 41.42 with 4.6\% performance gain while only requiring 10\% fewer GFLOPs and 25\% fewer parameters in the semantic branch. Furthermore, a real-world test via our designed compact wearable panoptic segmentation system, indicates that our system based on the Panoptic Lintention Net accomplishes a relatively stable and exceptionally remarkable panoptic segmentation in real-world scenes.
Panoptic segmentation unifies semantic segmentation and instance segmentation which has been attracting increasing attention in recent years. However, most existing research was conducted under a supervised learning setup whereas unsupervised domain adaptive panoptic segmentation which is critical in different tasks and applications is largely neglected. We design a domain adaptive panoptic segmentation network that exploits inter-style consistency and inter-task regularization for optimal domain adaptive panoptic segmentation. The inter-style consistency leverages geometric invariance across the same image of the different styles which fabricates certain self-supervisions to guide the network to learn domain-invariant features. The inter-task regularization exploits the complementary nature of instance segmentation and semantic segmentation and uses it as a constraint for better feature alignment across domains. Extensive experiments over multiple domain adaptive panoptic segmentation tasks (e.g., synthetic-to-real and real-to-real) show that our proposed network achieves superior segmentation performance as compared with the state-of-the-art.
How do we build a general and broad object detection system? We use all labels of all concepts ever annotated. These labels span diverse datasets with potentially inconsistent taxonomies. In this paper, we present a simple method for training a unified detector on multiple large-scale datasets. We use dataset-specific training protocols and losses, but share a common detection architecture with dataset-specific outputs. We show how to automatically integrate these dataset-specific outputs into a common semantic taxonomy. In contrast to prior work, our approach does not require manual taxonomy reconciliation. Experiments show our learned taxonomy outperforms a expert-designed taxonomy in all datasets. Our multi-dataset detector performs as well as dataset-specific models on each training domain, and can generalize to new unseen dataset without fine-tuning on them. Code is available at https://github.com/xingyizhou/UniDet.
Carton detection is an important technique in the automatic logistics system and can be applied to many applications such as the stacking and unstacking of cartons, the unloading of cartons in the containers. However, there is no public large-scale carton dataset for the research community to train and evaluate the carton detection models up to now, which hinders the development of carton detection. In this paper, we present a large-scale carton dataset named Stacked Carton Dataset(SCD) with the goal of advancing the state-of-the-art in carton detection. Images are collected from the internet and several warehourses, and objects are labeled using per-instance segmentation for precise localization. There are totally 250,000 instance masks from 16,136 images. In addition, we design a carton detector based on RetinaNet by embedding Offset Prediction between Classification and Localization module(OPCL) and Boundary Guided Supervision module(BGS). OPCL alleviates the imbalance problem between classification and localization quality which boosts AP by 3.1% - 4.7% on SCD while BGS guides the detector to pay more attention to boundary information of cartons and decouple repeated carton textures. To demonstrate the generalization of OPCL to other datasets, we conduct extensive experiments on MS COCO and PASCAL VOC. The improvement of AP on MS COCO and PASCAL VOC is 1.8% - 2.2% and 3.4% - 4.3% respectively.
Recent methods for long-tailed instance segmentation still struggle on rare object classes with few training data. We propose a simple yet effective method, Feature Augmentation and Sampling Adaptation (FASA), that addresses the data scarcity issue by augmenting the feature space especially for rare classes. Both the Feature Augmentation (FA) and feature sampling components are adaptive to the actual training status -- FA is informed by the feature mean and variance of observed real samples from past iterations, and we sample the generated virtual features in a loss-adapted manner to avoid over-fitting. FASA does not require any elaborate loss design, and removes the need for inter-class transfer learning that often involves large cost and manually-defined head/tail class groups. We show FASA is a fast, generic method that can be easily plugged into standard or long-tailed segmentation frameworks, with consistent performance gains and little added cost. FASA is also applicable to other tasks like long-tailed classification with state-of-the-art performance.
Although using convolutional neural networks (CNNs) as backbones achieves great successes in computer vision, this work investigates a simple backbone network useful for many dense prediction tasks without convolutions. Unlike the recently-proposed Transformer model (e.g., ViT) that is specially designed for image classification, we propose Pyramid Vision Transformer~(PVT), which overcomes the difficulties of porting Transformer to various dense prediction tasks. PVT has several merits compared to prior arts. (1) Different from ViT that typically has low-resolution outputs and high computational and memory cost, PVT can be not only trained on dense partitions of the image to achieve high output resolution, which is important for dense predictions but also using a progressive shrinking pyramid to reduce computations of large feature maps. (2) PVT inherits the advantages from both CNN and Transformer, making it a unified backbone in various vision tasks without convolutions by simply replacing CNN backbones. (3) We validate PVT by conducting extensive experiments, showing that it boosts the performance of many downstream tasks, e.g., object detection, semantic, and instance segmentation. For example, with a comparable number of parameters, RetinaNet+PVT achieves 40.4 AP on the COCO dataset, surpassing RetinNet+ResNet50 (36.3 AP) by 4.1 absolute AP. We hope PVT could serve as an alternative and useful backbone for pixel-level predictions and facilitate future researches. Code is available at https://github.com/whai362/PVT.
Large datasets' availability is catalyzing a rapid expansion of deep learning in general and computer vision in particular. At the same time, in many domains, a sufficient amount of training data is lacking, which may become an obstacle to the practical application of computer vision techniques. This paper challenges small and imbalanced datasets based on the example of a plant phenomics domain. We introduce an image augmentation framework, which enables us to extremely enlarge the number of training samples while providing the data for such tasks as object detection, semantic segmentation, instance segmentation, object counting, image denoising, and classification. We prove that our augmentation method increases model performance when only a few training samples are available. In our experiment, we use the DeepLabV3 model on semantic segmentation tasks with Arabidopsis and Nicotiana tabacum image dataset. The obtained result shows a 9% relative increase in model performance compared to the basic image augmentation techniques.
Cell segmentation and tracking in microscopy images are of great significance to new discoveries in biology and medicine. In this study, we propose a novel approach to combine cell segmentation and cell tracking into a unified end-to-end deep learning based framework, where cell detection and segmentation are performed with a current instance segmentation pipeline and cell tracking is implemented by integrating Siamese Network with the pipeline. Besides, tracking performance is improved by incorporating spatial information into the network and fusing spatial and visual prediction. Our approach was evaluated on the DeepCell benchmark dataset. Despite being simple and efficient, our method outperforms state-of-the-art algorithms in terms of both cell segmentation and cell tracking accuracies.
Instance segmentation aims to locate targets in the image and segment each target area at pixel level, which is one of the most important tasks in computer vision. Mask R-CNN is a classic method of instance segmentation, but we find that its predicted masks are unclear and inaccurate near contours. To cope with this problem, we draw on the idea of contour matching based on distance transformation image and propose a novel loss function, called contour loss. Contour loss is designed to specifically optimize the contour parts of the predicted masks, thus can assure more accurate instance segmentation. In order to make the proposed contour loss to be jointly trained under modern neural network frameworks, we design a differentiable k-step distance transformation image calculation module, which can approximately compute truncated distance transformation images of the predicted mask and corresponding ground-truth mask online. The proposed contour loss can be integrated into existing instance segmentation methods such as Mask R-CNN, and combined with their original loss functions without modification of the inference network structures, thus has strong versatility. Experimental results on COCO show that contour loss is effective, which can further improve instance segmentation performances.
Large labeled data sets are one of the essential basics of modern deep learning techniques. Therefore, there is an increasing need for tools that allow to label large amounts of data as intuitively as possible. In this paper, we introduce SALT, a tool to semi-automatically annotate RGB-D video sequences to generate 3D bounding boxes for full six Degrees of Freedom (DoF) object poses, as well as pixel-level instance segmentation masks for both RGB and depth. Besides bounding box propagation through various interpolation techniques, as well as algorithmically guided instance segmentation, our pipeline also provides built-in pre-processing functionalities to facilitate the data set creation process. By making full use of SALT, annotation time can be reduced by a factor of up to 33.95 for bounding box creation and 8.55 for RGB segmentation without compromising the quality of the automatically generated ground truth.
Many objects do not appear frequently enough in complex scenes (e.g., certain handbags in living rooms) for training an accurate object detector, but are often found frequently by themselves (e.g., in product images). Yet, these object-centric images are not effectively leveraged for improving object detection in scene-centric images. In this paper, we propose Mosaic of Object-centric images as Scene-centric images (MosaicOS), a simple and novel framework that is surprisingly effective at tackling the challenges of long-tailed object detection. Keys to our approach are three-fold: (i) pseudo scene-centric image construction from object-centric images for mitigating domain differences, (ii) high-quality bounding box imputation using the object-centric images' class labels, and (iii) a multi-stage training procedure. On LVIS object detection (and instance segmentation), MosaicOS leads to a massive 60% (and 23%) relative improvement in average precision for rare object categories. We also show that our framework can be compatibly used with other existing approaches to achieve even further gains. Our pre-trained models are publicly available at https://github.com/czhang0528/MosaicOS/.
We present lambda layers -- an alternative framework to self-attention -- for capturing long-range interactions between an input and structured contextual information (e.g. a pixel surrounded by other pixels). Lambda layers capture such interactions by transforming available contexts into linear functions, termed lambdas, and applying these linear functions to each input separately. Similar to linear attention, lambda layers bypass expensive attention maps, but in contrast, they model both content and position-based interactions which enables their application to large structured inputs such as images. The resulting neural network architectures, LambdaNetworks, significantly outperform their convolutional and attentional counterparts on ImageNet classification, COCO object detection and COCO instance segmentation, while being more computationally efficient. Additionally, we design LambdaResNets, a family of hybrid architectures across different scales, that considerably improves the speed-accuracy tradeoff of image classification models. LambdaResNets reach excellent accuracies on ImageNet while being 3.2 - 4.4x faster than the popular EfficientNets on modern machine learning accelerators. When training with an additional 130M pseudo-labeled images, LambdaResNets achieve up to a 9.5x speed-up over the corresponding EfficientNet checkpoints.
Panoptic segmentation of point clouds is a crucial task that enables autonomous vehicles to comprehend their vicinity using their highly accurate and reliable LiDAR sensors. Existing top-down approaches tackle this problem by either combining independent task-specific networks or translating methods from the image domain ignoring the intricacies of LiDAR data and thus often resulting in sub-optimal performance. In this paper, we present the novel top-down Efficient LiDAR Panoptic Segmentation (EfficientLPS) architecture that addresses multiple challenges in segmenting LiDAR point clouds including distance-dependent sparsity, severe occlusions, large scale-variations, and re-projection errors. EfficientLPS comprises of a novel shared backbone that encodes with strengthened geometric transformation modeling capacity and aggregates semantically rich range-aware multi-scale features. It incorporates new scale-invariant semantic and instance segmentation heads along with the panoptic fusion module which is supervised by our proposed panoptic periphery loss function. Additionally, we formulate a regularized pseudo labeling framework to further improve the performance of EfficientLPS by training on unlabelled data. We benchmark our proposed model on two large-scale LiDAR datasets: nuScenes, for which we also provide ground truth annotations, and SemanticKITTI. Notably, EfficientLPS sets the new state-of-the-art on both these datasets.
Instance segmentation has gained recently huge attention in various computer vision applications. It aims at providing different IDs to different object of the scene, even if they belong to the same class. This is useful in various scenarios, especially in occlusions. Instance segmentation is usually performed as a two-stage pipeline. First, an object is detected, then semantic segmentation within the detected box area. This process involves costly up-sampling, especially for the segmentation part. Moreover, for some applications, such as LiDAR point clouds and aerial object detection, it is often required to predict oriented boxes, which add extra complexity to the two-stage pipeline. In this paper, we propose Insta-YOLO, a novel one-stage end-to-end deep learning model for real-time instance segmentation. The proposed model is inspired by the YOLO one-shot object detector, with the box regression loss is replaced with polynomial regression in the localization head. This modification enables us to skip the segmentation up-sampling decoder altogether and produces the instance segmentation contour from the polynomial output coefficients. In addition, this architecture is a natural fit for oriented objects. We evaluate our model on three datasets, namely, Carnva, Cityscapes and Airbus. The results show our model achieves competitive accuracy in terms of mAP with significant improvement in speed by 2x on GTX-1080 GPU.
Camera traps are a proven tool in biology and specifically biodiversity research. However, camera traps including depth estimation are not widely deployed, despite providing valuable context about the scene and facilitating the automation of previously laborious manual ecological methods. In this study, we propose an automated camera trap-based approach to detect and identify animals using depth estimation. To detect and identify individual animals, we propose a novel method D-Mask R-CNN for the so-called instance segmentation which is a deep learning-based technique to detect and delineate each distinct object of interest appearing in an image or a video clip. An experimental evaluation shows the benefit of the additional depth estimation in terms of improved average precision scores of the animal detection compared to the standard approach that relies just on the image information. This novel approach was also evaluated in terms of a proof-of-concept in a zoo scenario using an RGB-D camera trap.
Low light conditions in aerial images adversely affect the performance of several vision based applications. There is a need for methods that can efficiently remove the low light attributes and assist in the performance of key vision tasks. In this work, we propose a new method that is capable of enhancing the low light image in a self-supervised fashion, and sequentially apply detection and segmentation tasks in an end-to-end manner. The proposed method occupies a very small overhead in terms of memory and computational power over the original algorithm and delivers superior results. Additionally, we propose the generation of a new low light aerial dataset using GANs, which can be used to evaluate vision based networks for similar adverse conditions.
Mapping new and old buildings are of great significance for understanding socio-economic development in rural areas. In recent years, deep neural networks have achieved remarkable building segmentation results in high-resolution remote sensing images. However, the scarce training data and the varying geographical environments have posed challenges for scalable building segmentation. This study proposes a novel framework based on Mask R-CNN, named HTMask R-CNN, to extract new and old rural buildings even when the label is scarce. The framework adopts the result of single-object instance segmentation from the orthodox Mask R-CNN. Further, it classifies the rural buildings into new and old ones based on a dynamic grayscale threshold inferred from the result of a two-object instance segmentation task where training data is scarce. We found that the framework can extract more buildings and achieve a much higher mean Average Precision (mAP) than the orthodox Mask R-CNN model. We tested the novel framework's performance with increasing training data and found that it converged even when the training samples were limited. This framework's main contribution is to allow scalable segmentation by using significantly fewer training samples than traditional machine learning practices. That makes mapping China's new and old rural buildings viable.