Stefan Milz

Sambit Mohapatra, Senthil Yogamani, Varun Ravi Kumar et al.

LiDAR is crucial for robust 3D scene perception in autonomous driving. LiDAR perception has the largest body of literature after camera perception. However, multi-task learning across tasks like detection, segmentation, and motion estimation using LiDAR remains relatively unexplored, especially on automotive-grade embedded platforms. We present a real-time multi-task convolutional neural network for LiDAR-based object detection, semantics, and motion segmentation. The unified architecture comprises a shared encoder and task-specific decoders, enabling joint representation learning. We propose a novel Semantic Weighting and Guidance (SWAG) module to transfer semantic features for improved object detection selectively. Our heterogeneous training scheme combines diverse datasets and exploits complementary cues between tasks. The work provides the first embedded implementation unifying these key perception tasks from LiDAR point clouds achieving 3ms latency on the embedded NVIDIA Xavier platform. We achieve state-of-the-art results for two tasks, semantic and motion segmentation, and close to state-of-the-art performance for 3D object detection. By maximizing hardware efficiency and leveraging multi-task synergies, our method delivers an accurate and efficient solution tailored for real-world automated driving deployment. Qualitative results can be seen at https://youtu.be/H-hWRzv2lIY.

Jack Borer, Jeremy Tschirner, Florian Ölsner et al.

Automated driving systems use multi-modal sensor suites to ensure the reliable, redundant and robust perception of the operating domain, for example camera and LiDAR. An accurate extrinsic calibration is required to fuse the camera and LiDAR data into a common spatial reference frame required by high-level perception functions. Over the life of the vehicle the value of the extrinsic calibration can change due physical disturbances, introducing an error into the high-level perception functions. Therefore there is a need for continuous online extrinsic calibration algorithms which can automatically update the value of the camera-LiDAR calibration during the life of the vehicle using only sensor data. We propose using mutual information between the camera image's depth estimate, provided by commonly available monocular depth estimation networks, and the LiDAR pointcloud's geometric distance as a optimization metric for extrinsic calibration. Our method requires no calibration target, no ground truth training data and no expensive offline optimization. We demonstrate our algorithm's accuracy, precision, speed and self-diagnosis capability on the KITTI-360 data set.

Jack Borer, Jeremy Tschirner, Florian Ölsner et al.

Sensor fusion is vital for the safe and robust operation of autonomous vehicles. Accurate extrinsic sensor to sensor calibration is necessary to accurately fuse multiple sensor's data in a common spatial reference frame. In this paper, we propose a target free extrinsic calibration algorithm that requires no ground truth training data, artificially constrained motion trajectories, hand engineered features or offline optimization and that is accurate, precise and extremely robust to initialization error. Most current research on online camera-LiDAR extrinsic calibration requires ground truth training data which is impossible to capture at scale. We revisit analytical mutual information based methods first proposed in 2012 and demonstrate that geometric features provide a robust information metric for camera-LiDAR extrinsic calibration. We demonstrate our proposed improvement using the KITTI and KITTI-360 fisheye data set.

Sambit Mohapatra, Senthil Yogamani, Heinrich Gotzig et al.

3D object detection based on LiDAR point clouds is a crucial module in autonomous driving particularly for long range sensing. Most of the research is focused on achieving higher accuracy and these models are not optimized for deployment on embedded systems from the perspective of latency and power efficiency. For high speed driving scenarios, latency is a crucial parameter as it provides more time to react to dangerous situations. Typically a voxel or point-cloud based 3D convolution approach is utilized for this module. Firstly, they are inefficient on embedded platforms as they are not suitable for efficient parallelization. Secondly, they have a variable runtime due to level of sparsity of the scene which is against the determinism needed in a safety system. In this work, we aim to develop a very low latency algorithm with fixed runtime. We propose a novel semantic segmentation architecture as a single unified model for object center detection using key points, box predictions and orientation prediction using binned classification in a simpler Bird's Eye View (BEV) 2D representation. The proposed architecture can be trivially extended to include semantic segmentation classes like road without any additional computation. The proposed model has a latency of 4 ms on the embedded Nvidia Xavier platform. The model is 5X faster than other top accuracy models with a minimal accuracy degradation of 2% in Average Precision at IoU=0.5 on KITTI dataset.

Varun Ravi Kumar, Marvin Klingner, Senthil Yogamani et al.

A 360° perception of scene geometry is essential for automated driving, notably for parking and urban driving scenarios. Typically, it is achieved using surround-view fisheye cameras, focusing on the near-field area around the vehicle. The majority of current depth estimation approaches focus on employing just a single camera, which cannot be straightforwardly generalized to multiple cameras. The depth estimation model must be tested on a variety of cameras equipped to millions of cars with varying camera geometries. Even within a single car, intrinsics vary due to manufacturing tolerances. Deep learning models are sensitive to these changes, and it is practically infeasible to train and test on each camera variant. As a result, we present novel camera-geometry adaptive multi-scale convolutions which utilize the camera parameters as a conditional input, enabling the model to generalize to previously unseen fisheye cameras. Additionally, we improve the distance estimation by pairwise and patchwise vector-based self-attention encoder networks. We evaluate our approach on the Fisheye WoodScape surround-view dataset, significantly improving over previous approaches. We also show a generalization of our approach across different camera viewing angles and perform extensive experiments to support our contributions. To enable comparison with other approaches, we evaluate the front camera data on the KITTI dataset (pinhole camera images) and achieve state-of-the-art performance among self-supervised monocular methods. An overview video with qualitative results is provided at https://youtu.be/bmX0UcU9wtA. Baseline code and dataset will be made public.

Varun Ravi Kumar, Senthil Yogamani, Hazem Rashed et al.

Surround View fisheye cameras are commonly deployed in automated driving for 360° near-field sensing around the vehicle. This work presents a multi-task visual perception network on unrectified fisheye images to enable the vehicle to sense its surrounding environment. It consists of six primary tasks necessary for an autonomous driving system: depth estimation, visual odometry, semantic segmentation, motion segmentation, object detection, and lens soiling detection. We demonstrate that the jointly trained model performs better than the respective single task versions. Our multi-task model has a shared encoder providing a significant computational advantage and has synergized decoders where tasks support each other. We propose a novel camera geometry based adaptation mechanism to encode the fisheye distortion model both at training and inference. This was crucial to enable training on the WoodScape dataset, comprised of data from different parts of the world collected by 12 different cameras mounted on three different cars with different intrinsics and viewpoints. Given that bounding boxes is not a good representation for distorted fisheye images, we also extend object detection to use a polygon with non-uniformly sampled vertices. We additionally evaluate our model on standard automotive datasets, namely KITTI and Cityscapes. We obtain the state-of-the-art results on KITTI for depth estimation and pose estimation tasks and competitive performance on the other tasks. We perform extensive ablation studies on various architecture choices and task weighting methodologies. A short video at https://youtu.be/xbSjZ5OfPes provides qualitative results.

Varun Ravi Kumar, Marvin Klingner, Senthil Yogamani et al.

State-of-the-art self-supervised learning approaches for monocular depth estimation usually suffer from scale ambiguity. They do not generalize well when applied on distance estimation for complex projection models such as in fisheye and omnidirectional cameras. This paper introduces a novel multi-task learning strategy to improve self-supervised monocular distance estimation on fisheye and pinhole camera images. Our contribution to this work is threefold: Firstly, we introduce a novel distance estimation network architecture using a self-attention based encoder coupled with robust semantic feature guidance to the decoder that can be trained in a one-stage fashion. Secondly, we integrate a generalized robust loss function, which improves performance significantly while removing the need for hyperparameter tuning with the reprojection loss. Finally, we reduce the artifacts caused by dynamic objects violating static world assumptions using a semantic masking strategy. We significantly improve upon the RMSE of previous work on fisheye by 25% reduction in RMSE. As there is little work on fisheye cameras, we evaluated the proposed method on KITTI using a pinhole model. We achieved state-of-the-art performance among self-supervised methods without requiring an external scale estimation.

Varun Ravi Kumar, Senthil Yogamani, Markus Bach et al.

In classical computer vision, rectification is an integral part of multi-view depth estimation. It typically includes epipolar rectification and lens distortion correction. This process simplifies the depth estimation significantly, and thus it has been adopted in CNN approaches. However, rectification has several side effects, including a reduced field of view (FOV), resampling distortion, and sensitivity to calibration errors. The effects are particularly pronounced in case of significant distortion (e.g., wide-angle fisheye cameras). In this paper, we propose a generic scale-aware self-supervised pipeline for estimating depth, euclidean distance, and visual odometry from unrectified monocular videos. We demonstrate a similar level of precision on the unrectified KITTI dataset with barrel distortion comparable to the rectified KITTI dataset. The intuition being that the rectification step can be implicitly absorbed within the CNN model, which learns the distortion model without increasing complexity. Our approach does not suffer from a reduced field of view and avoids computational costs for rectification at inference time. To further illustrate the general applicability of the proposed framework, we apply it to wide-angle fisheye cameras with 190$^\circ$ horizontal field of view. The training framework UnRectDepthNet takes in the camera distortion model as an argument and adapts projection and unprojection functions accordingly. The proposed algorithm is evaluated further on the KITTI rectified dataset, and we achieve state-of-the-art results that improve upon our previous work FisheyeDistanceNet. Qualitative results on a distorted test scene video sequence indicate excellent performance https://youtu.be/K6pbx3bU4Ss.

Victor Vaquero, Kai Fischer, Francesc Moreno-Noguer et al.

Localization and Mapping is an essential component to enable Autonomous Vehicles navigation, and requires an accuracy exceeding that of commercial GPS-based systems. Current odometry and mapping algorithms are able to provide this accurate information. However, the lack of robustness of these algorithms against dynamic obstacles and environmental changes, even for short time periods, forces the generation of new maps on every session without taking advantage of previously obtained ones. In this paper we propose the use of a deep learning architecture to segment movable objects from 3D LiDAR point clouds in order to obtain longer-lasting 3D maps. This will in turn allow for better, faster and more accurate re-localization and trajectoy estimation on subsequent days. We show the effectiveness of our approach in a very dynamic and cluttered scenario, a supermarket parking lot. For that, we record several sequences on different days and compare localization errors with and without our movable objects segmentation method. Results show that we are able to accurately re-locate over a filtered map, consistently reducing trajectory errors between an average of 35.1% with respect to a non-filtered map version and of 47.9% with respect to a standalone map created on the current session.

Varun Ravi Kumar, Sandesh Athni Hiremath, Stefan Milz et al.

Fisheye cameras are commonly used in applications like autonomous driving and surveillance to provide a large field of view ($>180^{\circ}$). However, they come at the cost of strong non-linear distortions which require more complex algorithms. In this paper, we explore Euclidean distance estimation on fisheye cameras for automotive scenes. Obtaining accurate and dense depth supervision is difficult in practice, but self-supervised learning approaches show promising results and could potentially overcome the problem. We present a novel self-supervised scale-aware framework for learning Euclidean distance and ego-motion from raw monocular fisheye videos without applying rectification. While it is possible to perform piece-wise linear approximation of fisheye projection surface and apply standard rectilinear models, it has its own set of issues like re-sampling distortion and discontinuities in transition regions. To encourage further research in this area, we will release our dataset as part of the WoodScape project \cite{yogamani2019woodscape}. We further evaluated the proposed algorithm on the KITTI dataset and obtained state-of-the-art results comparable to other self-supervised monocular methods. Qualitative results on an unseen fisheye video demonstrate impressive performance https://youtu.be/Sgq1WzoOmXg.

Martin Simon, Karl Amende, Andrea Kraus et al.

Accurate detection of 3D objects is a fundamental problem in computer vision and has an enormous impact on autonomous cars, augmented/virtual reality and many applications in robotics. In this work we present a novel fusion of neural network based state-of-the-art 3D detector and visual semantic segmentation in the context of autonomous driving. Additionally, we introduce Scale-Rotation-Translation score (SRTs), a fast and highly parameterizable evaluation metric for comparison of object detections, which speeds up our inference time up to 20\% and halves training time. On top, we apply state-of-the-art online multi target feature tracking on the object measurements to further increase accuracy and robustness utilizing temporal information. Our experiments on KITTI show that we achieve same results as state-of-the-art in all related categories, while maintaining the performance and accuracy trade-off and still run in real-time. Furthermore, our model is the first one that fuses visual semantic with 3D object detection.

Ganesh Sistu, Isabelle Leang, Sumanth Chennupati et al.

Convolutional Neural Networks (CNNs) are successfully used for the important automotive visual perception tasks including object recognition, motion and depth estimation, visual SLAM, etc. However, these tasks are typically independently explored and modeled. In this paper, we propose a joint multi-task network design for learning several tasks simultaneously. Our main motivation is the computational efficiency achieved by sharing the expensive initial convolutional layers between all tasks. Indeed, the main bottleneck in automated driving systems is the limited processing power available on deployment hardware. There is also some evidence for other benefits in improving accuracy for some tasks and easing development effort. It also offers scalability to add more tasks leveraging existing features and achieving better generalization. We survey various CNN based solutions for visual perception tasks in automated driving. Then we propose a unified CNN model for the important tasks and discuss several advanced optimization and architecture design techniques to improve the baseline model. The paper is partly review and partly positional with demonstration of several preliminary results promising for future research. We first demonstrate results of multi-stream learning and auxiliary learning which are important ingredients to scale to a large multi-task model. Finally, we implement a two-stream three-task network which performs better in many cases compared to their corresponding single-task models, while maintaining network size.

Sambit Mohapatra, Heinrich Gotzig, Senthil Yogamani et al.

Neural networks have become the standard model for various computer vision tasks in automated driving including semantic segmentation, moving object detection, depth estimation, visual odometry, etc. The main flavors of neural networks which are used commonly are convolutional (CNN) and recurrent (RNN). In spite of rapid progress in embedded processors, power consumption and cost is still a bottleneck. Spiking Neural Networks (SNNs) are gradually progressing to achieve low-power event-driven hardware architecture which has a potential for high efficiency. In this paper, we explore the role of deep spiking neural networks (SNN) for automated driving applications. We provide an overview of progress on SNN and argue how it can be a good fit for automated driving applications.

Timo Sämann, Karl Amende, Stefan Milz et al.

The ability to perform semantic segmentation in real-time capable applications with limited hardware is of great importance. One such application is the interpretation of the visual bird's-eye view, which requires the semantic segmentation of the four omnidirectional camera images. In this paper, we present an efficient semantic segmentation that sets new standards in terms of runtime and hardware requirements. Our two main contributions are the decrease of the runtime by parallelizing the ArgMax layer and the reduction of hardware requirements by applying the channel pruning method to the ENet model.