Dan Levi

Hila Levi, Guy Heller, Dan Levi et al.

The task of open-vocabulary object-centric image retrieval involves the retrieval of images containing a specified object of interest, delineated by an open-set text query. As working on large image datasets becomes standard, solving this task efficiently has gained significant practical importance. Applications include targeted performance analysis of retrieved images using ad-hoc queries and hard example mining during training. Recent advancements in contrastive-based open vocabulary systems have yielded remarkable breakthroughs, facilitating large-scale open vocabulary image retrieval. However, these approaches use a single global embedding per image, thereby constraining the system's ability to retrieve images containing relatively small object instances. Alternatively, incorporating local embeddings from detection pipelines faces scalability challenges, making it unsuitable for retrieval from large databases. In this work, we present a simple yet effective approach to object-centric open-vocabulary image retrieval. Our approach aggregates dense embeddings extracted from CLIP into a compact representation, essentially combining the scalability of image retrieval pipelines with the object identification capabilities of dense detection methods. We show the effectiveness of our scheme to the task by achieving significantly better results than global feature approaches on three datasets, increasing accuracy by up to 15 mAP points. We further integrate our scheme into a large scale retrieval framework and demonstrate our method's advantages in terms of scalability and interpretability.

Michael Baltaxe, Dan Levi, Sagie Benaim

Monocular Metric Depth Estimation (MMDE) is essential for physically intelligent systems, yet accurate depth estimation for underrepresented classes in complex scenes remains a persistent challenge. To address this, we propose RAD, a retrieval-augmented framework that approximates the benefits of multi-view stereo by utilizing retrieved neighbors as structural geometric proxies. Our method first employs an uncertainty-aware retrieval mechanism to identify low-confidence regions in the input and retrieve RGB-D context samples containing semantically similar content. We then process both the input and retrieved context via a dual-stream network and fuse them using a matched cross-attention module, which transfers geometric information only at reliable point correspondences. Evaluations on NYU Depth v2, KITTI, and Cityscapes demonstrate that RAD significantly outperforms state-of-the-art baselines on underrepresented classes, reducing relative absolute error by 29.2% on NYU Depth v2, 13.3% on KITTI, and 7.2% on Cityscapes, while maintaining competitive performance on standard in-domain benchmarks.

Hila Levi, Guy Heller, Dan Levi

As working with large datasets becomes standard, the task of accurately retrieving images containing objects of interest by an open set textual query gains practical importance. The current leading approach utilizes a pre-trained CLIP model without any adaptation to the target domain, balancing accuracy and efficiency through additional post-processing. In this work, we propose FOR: Finetuning for Object-centric Open-vocabulary Image Retrieval, which allows finetuning on a target dataset using closed-set labels while keeping the visual-language association crucial for open vocabulary retrieval. FOR is based on two design elements: a specialized decoder variant of the CLIP head customized for the intended task, and its coupling within a multi-objective training framework. Together, these design choices result in a significant increase in accuracy, showcasing improvements of up to 8 mAP@50 points over SoTA across three datasets. Additionally, we demonstrate that FOR is also effective in a semi-supervised setting, achieving impressive results even when only a small portion of the dataset is labeled.

Michael Baltaxe, Tomer Pe'er, Dan Levi

Autonomous driving and advanced driver-assistance systems rely on a set of sensors and algorithms to perform the appropriate actions and provide alerts as a function of the driving scene. Typically, the sensors include color cameras, radar, lidar and ultrasonic sensors. Strikingly however, although light polarization is a fundamental property of light, it is seldom harnessed for perception tasks. In this work we analyze the potential for improvement in perception tasks when using an RGB-polarimetric camera, as compared to an RGB camera. We examine monocular depth estimation and free space detection during the middle of the day, when polarization is independent of subject heading, and show that a quantifiable improvement can be achieved for both of them using state-of-the-art deep neural networks, with a minimum of architectural changes. We also present a new dataset composed of RGB-polarimetric images, lidar scans, GNSS / IMU readings and free space segmentations that further supports developing perception algorithms that take advantage of light polarization.

Netalee Efrat, Max Bluvstein, Shaul Oron et al.

3D-LaneNet+ is a camera-based DNN method for anchor free 3D lane detection which is able to detect 3d lanes of any arbitrary topology such as splits, merges, as well as short and perpendicular lanes. We follow recently proposed 3D-LaneNet, and extend it to enable the detection of these previously unsupported lane topologies. Our output representation is an anchor free, semi-local tile representation that breaks down lanes into simple lane segments whose parameters can be learnt. In addition we learn, per lane instance, feature embedding that reasons for the global connectivity of locally detected segments to form full 3d lanes. This combination allows 3D-LaneNet+ to avoid using lane anchors, non-maximum suppression, and lane model fitting as in the original 3D-LaneNet. We demonstrate the efficacy of 3D-LaneNet+ using both synthetic and real world data. Results show significant improvement relative to the original 3D-LaneNet that can be attributed to better generalization to complex lane topologies, curvatures and surface geometries.

Noa Garnett, Roy Uziel, Netalee Efrat et al.

Accurate lane detection, a crucial enabler for autonomous driving, currently relies on obtaining a large and diverse labeled training dataset. In this work, we explore learning from abundant, randomly generated synthetic data, together with unlabeled or partially labeled target domain data, instead. Randomly generated synthetic data has the advantage of controlled variability in the lane geometry and lighting, but it is limited in terms of photo-realism. This poses the challenge of adapting models learned on the unrealistic synthetic domain to real images. To this end we develop a novel autoencoder-based approach that uses synthetic labels unaligned with particular images for adapting to target domain data. In addition, we explore existing domain adaptation approaches, such as image translation and self-supervision, and adjust them to the lane detection task. We test all approaches in the unsupervised domain adaptation setting in which no target domain labels are available and in the semi-supervised setting in which a small portion of the target images are labeled. In extensive experiments using three different datasets, we demonstrate the possibility to save costly target domain labeling efforts. For example, using our proposed autoencoder approach on the llamas and tuSimple lane datasets, we can almost recover the fully supervised accuracy with only 10% of the labeled data. In addition, our autoencoder approach outperforms all other methods in the semi-supervised domain adaptation scenario.

Netalee Efrat, Max Bluvstein, Noa Garnett et al.

We propose a novel camera-based DNN method for 3D lane detection with uncertainty estimation. Our method is based on a semi-local, BEV, tile representation that breaks down lanes into simple lane segments. It combines learning a parametric model for the segments along with a deep feature embedding that is then used to cluster segment together into full lanes. This combination allows our method to generalize to complex lane topologies, curvatures and surface geometries. Additionally, our method is the first to output a learning based uncertainty estimation for the lane detection task. The efficacy of our method is demonstrated in extensive experiments achieving state-of-the-art results for camera-based 3D lane detection, while also showing our ability to generalize to complex topologies, curvatures and road geometries as well as to different cameras. We also demonstrate how our uncertainty estimation aligns with the empirical error statistics indicating that it is well calibrated and truly reflects the detection noise.

Dan Levi, Liran Gispan, Niv Giladi et al.

Predicting not only the target but also an accurate measure of uncertainty is important for many machine learning applications and in particular safety-critical ones. In this work we study the calibration of uncertainty prediction for regression tasks which often arise in real-world systems. We show that the existing definition for calibration of a regression uncertainty [Kuleshov et al. 2018] has severe limitations in distinguishing informative from non-informative uncertainty predictions. We propose a new definition that escapes this caveat and an evaluation method using a simple histogram-based approach. Our method clusters examples with similar uncertainty prediction and compares the prediction with the empirical uncertainty on these examples. We also propose a simple, scaling-based calibration method that preforms as well as much more complex ones. We show results on both a synthetic, controlled problem and on the object detection bounding-box regression task using the COCO and KITTI datasets.

Noa Garnett, Rafi Cohen, Tomer Pe'er et al.

We introduce a network that directly predicts the 3D layout of lanes in a road scene from a single image. This work marks a first attempt to address this task with on-board sensing without assuming a known constant lane width or relying on pre-mapped environments. Our network architecture, 3D-LaneNet, applies two new concepts: intra-network inverse-perspective mapping (IPM) and anchor-based lane representation. The intra-network IPM projection facilitates a dual-representation information flow in both regular image-view and top-view. An anchor-per-column output representation enables our end-to-end approach which replaces common heuristics such as clustering and outlier rejection, casting lane estimation as an object detection problem. In addition, our approach explicitly handles complex situations such as lane merges and splits. Results are shown on two new 3D lane datasets, a synthetic and a real one. For comparison with existing methods, we test our approach on the image-only tuSimple lane detection benchmark, achieving performance competitive with state-of-the-art.

Oran Shayer, Dan Levi, Ethan Fetaya

Recent breakthroughs in computer vision make use of large deep neural networks, utilizing the substantial speedup offered by GPUs. For applications running on limited hardware, however, high precision real-time processing can still be a challenge. One approach to solving this problem is training networks with binary or ternary weights, thus removing the need to calculate multiplications and significantly reducing memory size. In this work, we introduce LR-nets (Local reparameterization networks), a new method for training neural networks with discrete weights using stochastic parameters. We show how a simple modification to the local reparameterization trick, previously used to train Gaussian distributed weights, enables the training of discrete weights. Using the proposed training we test both binary and ternary models on MNIST, CIFAR-10 and ImageNet benchmarks and reach state-of-the-art results on most experiments.

Zhiding Yu, Wende Zhang, B. V. K. Vijaya Kumar et al.

We propose a vision-based highway border detection algorithm using structured Hough voting. Our approach takes advantage of the geometric relationship between highway road borders and highway lane markings. It uses a strategy where a number of trained road border and lane marking detectors are triggered, followed by Hough voting to generate corresponding detection of the border and lane marking. Since the initially triggered detectors usually result in large number of positives, conventional frame-wise Hough voting is not able to always generate robust border and lane marking results. Therefore, we formulate this problem as a joint detection-and-tracking problem under the structured Hough voting model, where tracking refers to exploiting inter-frame structural information to stabilize the detection results. Both qualitative and quantitative evaluations show the superiority of the proposed structured Hough voting model over a number of baseline methods.