Home

Mapillary instance segmentation

Mapillary Now Able to Recognize and Label Objects in the

  1. We believe that having semantic segmentation on street-level photos will open up many exciting avenues moving forward. One immediate application for Mapillary is to use semantic segmentation to improve other computer vision tasks. In 3D reconstruction, for instance, the presence of moving objects can decrease the quality of the reconstruction
  2. Today we're announcing that Mapillary will publish four papers at CVPR this year. In this post, we're looking at the paper named Seamless Scene Segmentation, which, as a world-first, rolls out a new computer vision model that slashes up to 20% computing powers when teaching machines to distinguish between people, cars, and map data like traffic signs, together with its overall environment
  3. Boosting Semantic Segmentation Performance with NVIDIA and Amazon. The new NVIDIA Tesla V100 graphics processing units and TensorRT 3.0 library together with Amazon EC2 P3 instances make Mapillary's semantic segmentation models 27 times faster while using 81% less memory. Advanced algorithms for semantic segmentation demand a lot of.
  4. We leverage state-of-the-art instance segmentation results in combination with optical flow predictions, also trained on automatically harvested training data. Our second major contribution is MOTSNet - a deep learning, tracking-by-detection architecture for MOTS - deploying a novel mask-pooling layer for improved object association over time
  5. The Mapillary Vistas Panoptic Segmentation Task targets the full perception stack for scene segmentation in street-images. Panoptic segmentation addresses both stuff and thing classes, unifying the typically distinct semantic and instance segmentation tasks. For full details of this task please see the Mapillary Vistas Panoptic Segmentation.
  6. Seamless Scene Segmentation is a CNN-based architecture that can be trained end-to-end to predict a complete class- and instance-specific labeling for each pixel in an image. To tackle this task, also known as Panoptic Segmentation, we take advantage of a novel segmentation head that seamlessly integrates multi-scale features generated by a.
  7. The Instance-specific Image Segmentation task requires that each car, pedestrian, etc. is segmented and labelled individually. Please find the detailed rules for each task on our research page . Both tasks will be using the Mapillary Vistas Dataset that we recently released

Introducing Seamless Scene Segmentation - Mapillary Blo

  1. e tracking labels for all detected object instances, building upon state-of-the-art optical flow models [49], that in turn have.
  2. The Mapillary Vistas Dataset is a large-scale street-level image dataset containing 25,000 high-resolution images annotated into 66/124 object categories of which 37/70 classes are instance-specific labels (v.1.2 and v2.0, respectively). Annotation is performed in a dense and fine-grained style by using polygons for delineating individual objects
  3. Mapillary Vistas Instance Segmentation Task Robust Vision @ ECCV 2020 Organized by Mapillary The goal of this task is to provide instance-specific segmentation results for the 37 thing classes in MVD

Boosting Semantic Segmentation Performance - Mapillar

Mapillary Street-Level Sequences: A Dataset for Lifelong Place Recognition By Frederik Warburg, Soren Hauberg, Manuel López-Antequera, Pau Gargallo, Yubin Kuang, Javier Civera Conf. on Computer Vision and Pattern Recognition (CVPR) 2020 / June, 2020 Learning Multi-Object Tracking and Segmentation from Automatic Annotation instance segmentation. However, the instance aggregation or object-level removal is still needed for results. In this paper, we represent objects in a box-free pipeline, which generates the kernel for each object and produces results by convolving the detail-rich feature directly, with no need for object-level duplicates removal [15, 37] 2.2. Instance Segmentation For the instance segmentation model, we adopt a two-pass pipeline. For more details, please see our instance seg-mentation technical report. 2.3. Panoptic Segmentation When we get the result of the stuff segmentation and the result of the segmentation, we need to fuse to get the final result of the panoptic segmentation We incorporate a new semantic head that aggregates fine and contextual features coherently and a new variant of Mask R-CNN as the instance head. We also propose a novel panoptic fusion module that congruously integrates the output logits from both the heads of our EfficientPS architecture to yield the final panoptic segmentation output With the initial release of the Mapillary Vistas dataset to researchers in May 2017, we invited the computer vision research community to work on challenging, real-world street-level data for image understanding tasks including semantic- or instance-specific segmentation, and later also panoptic segmentation. Our idea for Vistas was to compile a representative set of images from the Mapillary.

This dataset also consists of instance-level urban semantic segmentation for 37 classes out of 66. Since the images on the Mapillary platform are collaboratively collected, they are from a variety of viewing angles, as is visible through this explorer; One can make submissions of algorithms on their dataset over here. Final Thought Mapillary Research research@mapillary.com Abstract The Mapillary Vistas Dataset is a novel, large-scale street-level image dataset containing 25000 high-resolution images annotated into 66 object categories with additional, instance-specific labels for 37 classes. Annota-tion is performed in a dense and fine-grained style by usin

We are pleased to contribute the Mapillary Vistas dataset to the Robust Vision Workshop @ ECCV 2020: Panoptic Segmentation Task. Panoptic segmentation addresses both stuff and thing classes, unifying the typically distinct semantic and instance segmentation tasks Detailed results on semantic segmentation and instance-specific semantic segmentation In Tab. 1 we report the per-category AP and AP@0.5 scores obtained for the Mapillary Vistas Dataset's instance-specific semantic segmentation task on both, validation set and test set for the Mask R-CNN variant of the UCenter team that won the LSUN challenge

The Mapillary Vistas Dataset is a novel, large-scale street-level image dataset containing 25000 high-resolution images annotated into 66 object categories with additional, instance-specific labels for 37 classes. Annotation is performed in a dense and fine-grained style by using polygons for delineating individual objects. Our dataset is 5× larger than the total amount of fine annotations. For instance segmentation, a Mask R-CNN type of architecture is used, while the semantic segmentation branch is augmented with a Pyramid Pooling Module. Results for this method are submitted to the COCO and Mapillary Joint Recognition Challenge 2018. Our approach achieves a PQ score of 17.6 on the Mapillary Vistas validation set and 27.2 on the. KITTI. The KITTI semantic segmentation dataset consists of 200 semantically annotated training images and of 200 test images. The total KITTI dataset is not only for semantic segmentation, it also includes dataset of 2D and 3D object detection, object tracking, road/lane detection, scene flow, depth evaluation, optical flow and semantic instance level segmentation How to reproduce the HRNet + OCR with Mapillary pretraining. SegFix can be used to improve the semantic/instance segmentation results of any existing approaches, e.g., HRNet, DeepLabv3, OCR, PointRend, MaskRCNN, without any re-training or fine-tuning. We have made the inference code and the offset files of our SegFix method

Mapillary Researc

In this report, we present our object detection/instance segmentation system, MegDetV2, which works in a two-pass fashion, first to detect instances then to obtain segmentation. Our baseline detector is mainly built on a new designed RPN, called RPN++. On the COCO-2019 detection/instance-segmentation test-dev dataset, our system achieves 61.0/53.1 mAP, which surpassed our 2018 winning results. These files are not needed for semantic and instance segmentation. Expected dataset structure for Mapillary Vistas: mapillary_vistas/ training/ images/ instances/ labels/ panoptic/ validation/ images/ instances/ labels/ panoptic/ No preprocessing is needed for Mapillary Vistas.. The semantic segmentation branch is the same as the typical design of any semantic segmentation model (e.g., DeepLab), while the instance segmentation branch is class-agnostic, involving a simple instance center regression

We present Adaptive Instance Selection network architecture for class-agnostic instance segmentation. Given an input image and a point ( x, y), it generates a mask for the object located at ( x, y).. The network adapts to the input point with a help of AdaIN layers, thus producing different masks for different objects on the same image Joint COCO and Mapillary Workshop at ICCV 2019: COCO Instance Segmentation Challenge Track. 10/06/2020 ∙ by Zeming Li, et al. ∙ 0 ∙ share . In this report, we present our object detection/instance segmentation system, MegDetV2, which works in a two-pass fashion, first to detect instances then to obtain segmentation

COCO + Mapillary ICCV 201

  1. For instance segmentation, a Mask R-CNN type of architecture is used, while the semantic segmentation branch is augmented with a Pyramid Pooling Module. Results for this method are submitted to the COCO and Mapillary Joint Recognition Challenge 2018
  2. For instance segmentation, a Mask R-CNN type of architecture is used, while the semantic segmentation branch is augmented with a Pyramid Pooling Module. Results for this method are submitted to the COCO and Mapillary Joint Recognition Challenge 2018. Our approach achieves a PQ score of 17.6 on the Mapillary Vistas validation set and 27.2 on the.
  3. Panoptic segmentation is a recently proposed task that features a challenging unification of semantic segmentation and instance segmentation. It provides a holistic solution to scene parsing by predicting both pixel-level classification and instance labels. To pursue a high performance evolving around the proposed metric Panoptic Quality (PQ.
  4. 3.2. Instance Segmentation Branch The instance segmentation branch follows the state-of-the-art Mask R-CNN [19] architecture, but with some mod-ifications described next. This branch is structured into a region proposal head and a region segmentation head. Region Proposal Head (RPH). The RPH introduces the notion of an anchor
  5. Instance segmentation: Instance segmentation recognizes and localizes object instances with high pixel-level accuracy. DeepLab tackles the instance segmentation by detecting the instances nothing but a group of semantic objects on top of segmentation prediction
  6. Panoptic segmentation [16] is the task of generating per-pixel, semantic labels for an image, together with object-specific segmentation masks. It is thus a combination of semantic segmentation and instance segmentation, i.e. two long-standing tasks in computer vision that have been tradi-tionally tackled separately. Due to its importance for.
  7. A new dataset for long tail object detection. @inproceedings{gupta2019lvis, title={{LVIS}: A Dataset for Large Vocabulary Instance Segmentation}, author={Gupta, Agrim and Dollar, Piotr and Girshick, Ross}, booktitle={Proceedings of the {IEEE} Conference on Computer Vision and Pattern Recognition}, year={2019}

GitHub - mapillary/seamseg: Seamless Scene Segmentatio

Mapillary Stuff Method Stuff mIoU(%) Baseline(Res50) 56.3 +Residual L2 Loss 58.0 +Multiscale Testing 58.7 +Large Model 62.4 +3 Model Ensemble 62.8 Evaluation of semantic segmentation on the Val datase K-Net: Towards Unified Image Segmentation. open-mmlab/mmdetection • • 28 Jun 2021 The framework, named K-Net, segments both instances and semantic categories consistently by a group of learnable kernels, where each kernel is responsible for generating a mask for either a potential instance or a stuff class The architecture consists of a ResNet-50 feature extractor shared by the semantic segmentation and instance segmentation branch. For instance segmentation, a Mask R-CNN type of architecture is used, while the semantic segmentation branch is augmented with a Pyramid Pooling Module. Results for this method are submitted to the COCO and Mapillary. In the COCO and Mapillary Joint Recognition Challenge Workshop with ICCV 2019, the COCO Dense Pose challenge winner and almost all the COCO keypoint detection challenge participants adopted the HRNet. The OpenImage instance segmentation challenge winner (ICCV 2019) also used the HRNet The semantic segmentation task uses all the publicly available Mapillary Vistas Research edition v1.2 images (18.000 train + 2.000 val). The main performance metric used is mean Intersection-over-Union (mIoU) computed on 65 valid object categories. This CodaLab evaluation server provides a platform to measure performance on the val and test sets

In this work, we propose a single deep neural network for panoptic segmentation, for which the goal is to provide each individual pixel of an input image with a class label, as in semantic segmentation, as well as a unique identifier for specific objects in an image, following instance segmentation. Our network makes joint semantic and instance segmentation predictions and combines these to. EfficientPS is currently ranked #1 for panoptic segmentation on standard benchmark datasets such as Cityscapes, KITTI, Mapillary Vistas, and IDD. Additionally, EfficientPS is also ranked #2 on the Cityscapes semantic segmentation benchmark as well as #2 on the Cityscapes instance segmentation benchmark, among the published methods

The Mapillary Vistas Dataset is a novel, large-scale street-level image dataset containing 25,000 high-resolution images annotated into 66 object categories with additional, instance-specific labels for 37 classes. Annotation is performed in a dense and fine-grained style by using polygons for delineating individual objects Semantic Image Segmentation - Source: Sample from the Mapillary Vistas Dataset Applications of Image Segmentation. Image segmentation plays a central role in a broad range of real-world computer vision applications, including road sign detection, biology, the evaluation of construction materials, or video surveillance.Also, autonomous vehicles and Advanced Driver Assistance Systems (ADAS. Panoptic segmentation is a recently introduced scene understanding problem (Kirillov et al 2019b) that unifies the tasks of semantic segmentation and instance segmentation.There are numerous methods that have been proposed for each of these sub-tasks, however only a handful of approaches have been introduced to tackle this coherent scene understanding problem of panoptic segmentation mpirun.real detected that one or more processes exited with non-zero status, thus causing the job to be terminated. The first process to do so was for classi cation, and on three datasets (COCO [56], Mapillary Vistas [62], and Cityscapes [22]) for panoptic segmentation [45], instance segmentation, and se-mantic segmentation. In particular, on ImageNet, we build an Axial-ResNet by replacing the 3 3 convolution in all residual blocks [31] with our position

Pushing the Limits of Scene Understanding - Mapillary Blo

In this work, we introduce Panoptic-DeepLab, a simple, strong, and fast system for panoptic segmentation, aiming to establish a solid baseline for bottom-up methods that can achieve comparable performance of two-stage methods while yielding fast inference speed. In particular, Panoptic-DeepLab adopts the dual-ASPP and dual-decoder structures specific to semantic, and instance segmentation. Multi-object tracking and segmentation Mapillary-scale data for learning single-image depth estimation, extracted from multiple cameras and all around the globe, using SfM SOTA recognition algorithms for automatically mining training data is beneficial for MOTS. Even possible to outperform methods based on manually annotated dat

Mapillar

Abstract 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 is also ranked second for the semantic segmentation task as well as the instance segmentation task on the Cityscapes benchmark with a mIoU score of 84:2% and an AP of 39:1% respectively. On the Mapillary Vistas dataset, our single Effi-cientPS model achieves a PQ score of 40:5% on the valida-tion set, thereby outperforming all the existing. Panoptic Segmentation with a Joint Semantic and Instance Segmentation Network. arXiv (2018). Daan de Geus, Panagiotis Meletis, Gijs Dubbelman. Single Network Panoptic Segmentation for Street Scene Understanding. arXiv (2019). David Owen, Ping-Lin Chang. Detecting Reflections by Combining Semantic and Instance Segmentation

previously trained on the COCO dataset and fine-tune it for Mapillary images to obtain instance-level segmentation outputs. 2 Dataset The Mapillary Vistas dataset [3] contains 20,000 high-resolution street-level images on multiple locations around the world. 37 object categories are labeled with pixel-wise instance-level annotations The first to perform instance segmentation and on imagery by fusing images and disparity information to regress object masks. We collect High-Quality Driving Stereo (HQDS) with f x b 4 times larger than Mapillary Cityscapes KITTI HQDS Stereo X Resolution (megapixels) <0.5 7.99 2.09 0.71 3.15 Stereo Pairs # 2.7K 0.2K 6K Baseline (m) 0.2 0.5. Important variants • Partial semantic segmentation • some pixels unlabelled • Thing segmentation • label things • count nouns (car, person, dog) • Stuff segmentation • label stuff • mass nouns (grass, sky, water) • Panoptic segmentation • each pixel gets a label • each instance of a count noun gets a different label (person-a, etc Unifying Semantic and Instance Segmentation Semantic Segmentation • per-pixel annotation • simple accuracy measure • instances indistinguishable Object Detection/Seg • each object detected and segmented separately • stuff is not segmented Mapillary Vistas + + grass.

A Dataset for Lane Instance Segmentation in Urban Environments 3 average annotation time per image is much lower. However, our provided classes are different, since we focus on lane instances (and thus ignore other semantic segmentation classes like vehicle, building, person, etc.). Furthermore, our dat The solution presented in this article employs the Mask R-CNN algorithm for instance segmentation, which is an evolution of the Faster R-CNN (designed for object detection). I will not elaborate. Current state-of-the-art seamless segmentation networks 11 can be trained to identify billboards using the Mapillary Vistas Dataset for semantic understanding of street scenes For instance, in.

CodaLab - Competitions - Mapillar

  1. Instance segmentation performance as a function of mask size. Both Mapillary and GSV panoramas showed suitable for urban tree mapping. We accurately geolocated trees using a monocular depth estimation algorithm and triangulation that requires no additional contextual information. The geolocation of street trees with a mean accuracy of.
  2. EfficientPS: Efficient Panoptic Segmentation. Understanding the scene in which an autonomous robot operates is critical for its competent functioning. Such scene comprehension necessitates recognizing instances of traffic participants along with general scene semantics which can be effectively addressed by the panoptic segmentation task
  3. Introduction. The goal in panoptic segmentation is to perform a unified segmentation task. In order to do so, let's first understand few basic concepts. A thing is a countable object such as people, car, etc, thus it's a category having instance-level annotation. The stuff is amorphous region of similar texture such as road, sky, etc, thus.

As our videos are in a different domain, we provide instance segmentation annotations as well to compare the domain shift relative by different datasets. It can be expensive and laborious to obtain full pixel-level segmentation. Fortunately, with our own labeling tool, the labeling cost could be reduced by 50% semantic-, instance- and panoptic segmentation algorithms [5,18,22,31] has been greatly improved within a few years only. Yet, metric-accurate, large-scale, natural image datasets are still to come for the task of monocular depth estimation, most likely because they cannot be collected with commodity hardware in a straightfor-ward way On the memory-demanding task of semantic segmentation, we report competitive results for COCO-Stuff and set new state-of-the-art results for Cityscapes and Mapillary Vistas. Example of residual block with identity mapping. Left: Implementation with standard BN and in-place activation layers, which requires storing 6 buffers for the backward pass Mapillary will be able to use these tools as a foundation for processing their photos while also Currently, the semantic segmentation process will classify an object with a minimum of 64 pixels. But, even when combined with a specific For instance, a transit network model can be seriously altered by a bike.

Mapillary [22]) and pedestrians (Caltech Pedestrians [4]). We review the most closely related datasets below. COCO [18] is the most popular instance segmentation benchmark for common objects. It contains 80 categories that are pairwise distinct. There are a total of 118k train-ing images, 5k validation images, and 41k test images. Al More recently, Mapillary benchmarked TensorRT 3.0 running on Tesla V100 GPUs via the Amazon Web Services EC2 P3 instance. The result was a 27x speed-up of HD segmentation while reducing memory demands by 81 percent. Standard segmentation was boosted by 18x, with a 74 percent memory reduction • Partial semantic segmentation • some pixels unlabelled • Thing segmentation • label things • count nouns (car, person, dog) • Stuff segmentation • label stuff • mass nouns (grass, sky, water) • Panoptic segmentation • each pixel gets a label • each instance of a count noun gets a different label (person. approaches are designed to generate instances without assis-tance of object boxes [10,4,45,46]. Recently, AdaptIS [40] and CondInst [42] are proposed to utilize point-proposal for instance segmentation. However, the instance aggregation or object-level removal is still needed for results. In this paper, we represent objects in a box-free. Abstract. Obtaining precise instance segmentation masks is of high importance in many modern applications such as robotic manipulation and autonomous driving. Currently, many state of the art models are based on the Mask R-CNN framework which, while very powerful, outputs masks at low resolutions which could result in imprecise boundaries

Mapillary Research - Seamless Scene Segmentatio

Joint COCO and Mapillary Workshop at ICCV 2019: COCO Instance Segmentation Challenge Track Zeming Li, Yuchen Ma, Yukang Chen, Xiangyu Zhang, Jian Sun. 1st place in the Micro-soft COCO Instance Segmentation Challenge In ICCV 2019 Workshop for instance segmentation). Top-performing instance seg-mentation methods such as Mask R-CNN rely on ROI op-erations (typically ROIPool or ROIAlign) to obtain the fi-nal instance masks. In contrast, we propose to solve in-stance segmentation from a new perspective. Instead of using instance-wise ROIs as inputs to a network of fixe Mapillary (mapillary.com) is a service for sharing geotagged photos developed by a Swedish startup, that was sold to Facebook in June 2020. Its creators want to represent the whole world (not only streets) with photos .They believe that for covering all interesting places in the world, there needs to be an independent, crowd-sourced project and a systematic approach to cover interesting areas The videos comes with GPU/IMU data for trajectory information. They are manually tagged with weather, time of the day, and scene types. We also labeled bounding boxes of all the objects on the road, lane markings, drivable areas and detailed full-frame instance segmentation

instance-level segmentation is the Mask R-CNN model. The model also uses convolution extensively and has been shown to work well for segmentation task. One of the future goal would be to either adopt or impelment this model and test on the Mapillary data set. As seen in the current result, objects of small or narrow dimensions are not well. The instance Selection Network, in the end, takes the provided inputs and tries to generate a binary mask that will give the object segmentation. The evaluations that researchers conducted show that the method achieves state-of-the-art results on the Cityscapes and Mapillary datasets even without pre-training Fast Semantic Segmentation. This respository aims to provide accurate real-time semantic segmentation code for mobile devices in PyTorch, with pretrained weights on Cityscapes. This can be used for efficient segmentation on a variety of real-world street images, including datasets like Mapillary Vistas, KITTI, and CamVid Unified view of semantic- and instance-level segmentation tasks. Support major semantic segmentation datasets: ADE20K, Cityscapes, COCO-Stuff, Mapillary Vistas. Support ALL Detectron2 models. Installation. See installation instructions. Getting Started. See Preparing Datasets for MaskFormer. See Getting Started with MaskFormer. Model Zoo and.

Instance Segmentation Tracking Bounding Box Tracking Panoptic Segmentation Drivable Area Lane & Tagging Sunny City Street Daytime. Pascal COCO Mapillary Waymo ArgoversenuScenes Youtube-BB BDD100K Images 10K 328k 25K - - - - - Videos - - - 2K 113 1K 240K 100 Seamless Scene Segmentation. Abstract: In this work we introduce a novel, CNN-based architecture that can be trained end-to-end to deliver seamless scene segmentation results. Our goal is to predict consistent semantic segmentation and detection results by means of a panoptic output format, going beyond the simple combination of independently. We present Panoptic-DeepLab, a bottom-up and single-shot approach for panoptic segmentation. Our Panoptic-DeepLab is conceptually simple and delivers state-of-the-art results. In particular, we adopt the dual-ASPP and dual-decoder structures specific to semantic, and instance segmentation, respectively. The semantic segmentation branch is the same as the typical design of any semantic.

EfficientPS: Efficient Panoptic Segmentation Papers With

In this work, we propose a single deep neural network for panoptic segmentation, for which the goal is to provide each individual pixel of an input image with a class label, as in semantic segmentation, as well as a unique identifier for specific objects in an image, following instance segmentation Mapping urban trees with deep learning and street-level imagery. Creator. Lumnitz, Stefanie. Publisher. University of British Columbia. Date Issued. 2019. Description. Planning and managing urban trees and forests for livable cities remains an outstanding challenge worldwide owing to scarce information on their spatial distribution, structure.

Upgrading to Vistas 2

Besides making street-level imagery available to anyone, Mapillary also uses computer vision to automatically extract map data from images. You can use two types of data. 1) Object detections are instances of different objects that have been detected in images. Since Mapillary images are geotagged, you can get a dataset of image locations for a. The Robust Vision Challenge 2018 was a full day event held in conjunction with CVPR 2018 in Salt Lake City. Our workshop comprised talks by the winning teams of each challenge as well as three invited keynote talks by renowned experts in the field

Unified view of semantic- and instance-level segmentation tasks. Support major semantic segmentation datasets: ADE20K, Cityscapes, COCO-Stuff, Mapillary Vistas. Support ALL Detectron2 models. Same exact model, loss, and training procedur Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation. Click To Get Model/Code. In this work, we introduce Panoptic-DeepLab, a simple, strong, and fast system for panoptic segmentation, aiming to establish a solid baseline for bottom-up methods that can achieve comparable performance of two-stage methods while yielding fast inference speed. In particular. In particular, our model outperforms all existing stand-alone self-attention models on ImageNet. Our Axial-DeepLab improves 2.8% PQ over bottom-up state-of-the-art on COCO test-dev. This previous state-of-the-art is attained by our small variant that is 3.8\times parameter-efficient and 27\times computation-efficient

Panoptic segmentation unifies the typically distinct tasks of semantic segmentation (assign a class label to each pixel) and instance segmentation (detect and segment each object instance). The proposed task requires generating a coherent scene segmentation that is rich and complete, an important step toward real-world vision systems Hi, I'm Frederik Warburg . I'm a PhD Student in Uncertainty-Aware Deep Learning for Autonomous Driving at the Technical Unversity of Denmark, supervised by Søren Hauberg, Javier Civera and Søren K. S. Gregersen. My research interests are in the cross section between Computer Vision and Deep Learning, especially lifelong place recogntion, uncertainty quantification, 3D reconstruction, visual. The team from China has won the Microsoft COCO Challenge for two consecutive years, and this article will analyze the real secrets of their success. Microsoft Common Objects in Context (COCO) is Training of Convolutional Networks on Multiple Heterogeneous Datasets for Street Scene Semantic Segmentation Conference IEEE Intelligent Vehicles, 2018 , ISBN: 978-1-5386-4452-2 . Links | BibTeX | Altmetri

Semantic Segmentation Datasets for Autonomous DrivingCOCO + Mapillary 2018 | ECCV 2018DeeperLab Project