Martial Hebert
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Martial Hebert has a total of 57 co-authors. Here are the top co-authors by number of shared publications:
- Alexei A. Efros
UC Berkeley
2 shared publications Anthony Stentz
Carnegie Mellon University
2 shared publications- Bowei Chen
Carnegie Mellon University
2 shared publications - Jean-François Lalonde
Université Laval
2 shared publications Nicolas Vandapel
Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213, USA
2 shared publications- Srinivasa G. Narasimhan
Carnegie Mellon University
2 shared publications
Latest Publications
- 18 June 2026
- Published by Cassyni
Advances in intelligent machines and autonomy and your thoughts on research needs and gaps
The panel will discuss diverse topics and issues related to the development of autonomous capabilities, their enabling technologies – including artificial intelligence (AI), and their application to aerospace and robotic systems more generally. Some specific topics include: - Constraints on currently fielded autonomous capabilities; - Roles for contemporary AI in design, development, and operations of robotic vehicle systems; - Limitations on the use of AI in safety-critical applications; - Establishing trust in AI methods; - Issues with the certification of autonomous capabilities using established methods and processes; and, - The evolutionary traits needed for AI methods to render trusted, acceptable judgment in the performance of complex missions by autonomous robotic systems; … to name a few. The audience will be afforded a live question-answer period at the conclusion of panel presentations.
- 18 June 2026
- Published by Cassyni
Advances in intelligent machines and autonomy and your thoughts on research needs and gaps
The panel will discuss diverse topics and issues related to the development of autonomous capabilities, their enabling technologies – including artificial intelligence (AI), and their application to aerospace and robotic systems more generally. Some specific topics include: - Constraints on currently fielded autonomous capabilities; - Roles for contemporary AI in design, development, and operations of robotic vehicle systems; - Limitations on the use of AI in safety-critical applications; - Establishing trust in AI methods; - Issues with the certification of autonomous capabilities using established methods and processes; and, - The evolutionary traits needed for AI methods to render trusted, acceptable judgment in the performance of complex missions by autonomous robotic systems; … to name a few. The audience will be afforded a live question-answer period at the conclusion of panel presentations.
- 20 May 2026
- Published by Springer Nature
- p. 207-226
Geodesic Turnpikes for Robot Motion Planning
- Open Access
- 13 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-10
Separate-and-Enhance: Compositional Finetuning for Text-to-Image Diffusion Models
Despite recent significant strides achieved by diffusion-based Text-to-Image (T2I) models, current systems are still less capable of ensuring decent compositional generation aligned with text prompts, particularly for the multi-object generation. In this work, we first show the fundamental reasons for such misalignment by identifying issues related to low attention activation and mask overlaps. Then we propose a compositional finetuning framework with two novel objectives, the Separate loss and the Enhance loss, that reduce object mask overlaps and maximize attention scores, respectively. Unlike conventional test-time adaptation methods, our model, once finetuned on critical parameters, is able to directly perform inference given an arbitrary multi-object prompt, which enhances the scalability and generalizability. Through comprehensive evaluations, our model demonstrates superior performance in image realism, text-image alignment, and adaptability, significantly surpassing established baselines. Furthermore, we show that training our model with a diverse range of concepts enables it to generalize effectively to novel concepts, exhibiting enhanced performance compared to models trained on individual concept pairs.
- 13 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-10
Open AccessSeparate-and-Enhance: Compositional Finetuning for Text-to-Image Diffusion Models
Despite recent significant strides achieved by diffusion-based Text-to-Image (T2I) models, current systems are still less capable of ensuring decent compositional generation aligned with text prompts, particularly for the multi-object generation. In this work, we first show the fundamental reasons for such misalignment by identifying issues related to low attention activation and mask overlaps. Then we propose a compositional finetuning framework with two novel objectives, the Separate loss and the Enhance loss, that reduce object mask overlaps and maximize attention scores, respectively. Unlike conventional test-time adaptation methods, our model, once finetuned on critical parameters, is able to directly perform inference given an arbitrary multi-object prompt, which enhances the scalability and generalizability. Through comprehensive evaluations, our model demonstrates superior performance in image realism, text-image alignment, and adaptability, significantly surpassing established baselines. Furthermore, we show that training our model with a diverse range of concepts enables it to generalize effectively to novel concepts, exhibiting enhanced performance compared to models trained on individual concept pairs.
- Open Access
Thinking Like an Annotator: Generation of Dataset Labeling Instructions
Large-scale datasets are essential to modern day deep learning. Advocates argue that understanding these methods requires dataset transparency (e.g. "dataset curation, motivation, composition, collection process, etc..."). However, almost no one has suggested the release of the detailed definitions and visual category examples provided to annotators - information critical to understanding the structure of the annotations present in each dataset. These labels are at the heart of public datasets, yet few datasets include the instructions that were used to generate them. We introduce a new task, Labeling Instruction Generation, to address missing publicly available labeling instructions. In Labeling Instruction Generation, we take a reasonably annotated dataset and: 1) generate a set of examples that are visually representative of each category in the dataset; 2) provide a text label that corresponds to each of the examples. We introduce a framework that requires no model training to solve this task and includes a newly created rapid retrieval system that leverages a large, pre-trained vision and language model. This framework acts as a proxy to human annotators that can help to both generate a final labeling instruction set and evaluate its quality. Our framework generates multiple diverse visual and text representations of dataset categories. The optimized instruction set outperforms our strongest baseline across 5 folds by 7.06 mAP for NuImages and 12.9 mAP for COCO.
Open AccessThinking Like an Annotator: Generation of Dataset Labeling Instructions
Large-scale datasets are essential to modern day deep learning. Advocates argue that understanding these methods requires dataset transparency (e.g. "dataset curation, motivation, composition, collection process, etc..."). However, almost no one has suggested the release of the detailed definitions and visual category examples provided to annotators - information critical to understanding the structure of the annotations present in each dataset. These labels are at the heart of public datasets, yet few datasets include the instructions that were used to generate them. We introduce a new task, Labeling Instruction Generation, to address missing publicly available labeling instructions. In Labeling Instruction Generation, we take a reasonably annotated dataset and: 1) generate a set of examples that are visually representative of each category in the dataset; 2) provide a text label that corresponds to each of the examples. We introduce a framework that requires no model training to solve this task and includes a newly created rapid retrieval system that leverages a large, pre-trained vision and language model. This framework acts as a proxy to human annotators that can help to both generate a final labeling instruction set and evaluate its quality. Our framework generates multiple diverse visual and text representations of dataset categories. The optimized instruction set outperforms our strongest baseline across 5 folds by 7.06 mAP for NuImages and 12.9 mAP for COCO.
- Open Access
Deep Projective Rotation Estimation through Relative Supervision
Orientation estimation is the core to a variety of vision and robotics tasks such as camera and object pose estimation. Deep learning has offered a way to develop image-based orientation estimators; however, such estimators often require training on a large labeled dataset, which can be time-intensive to collect. In this work, we explore whether self-supervised learning from unlabeled data can be used to alleviate this issue. Specifically, we assume access to estimates of the relative orientation between neighboring poses, such that can be obtained via a local alignment method. While self-supervised learning has been used successfully for translational object keypoints, in this work, we show that naively applying relative supervision to the rotational group $SO(3)$ will often fail to converge due to the non-convexity of the rotational space. To tackle this challenge, we propose a new algorithm for self-supervised orientation estimation which utilizes Modified Rodrigues Parameters to stereographically project the closed manifold of $SO(3)$ to the open manifold of $\mathbb{R}^{3}$, allowing the optimization to be done in an open Euclidean space. We empirically validate the benefits of the proposed algorithm for rotational averaging problem in two settings: (1) direct optimization on rotation parameters, and (2) optimization of parameters of a convolutional neural network that predicts object orientations from images. In both settings, we demonstrate that our proposed algorithm is able to converge to a consistent relative orientation frame much faster than algorithms that purely operate in the $SO(3)$ space. Additional information can be found at https://sites.google.com/view/deep-projective-rotation/home .Open AccessDeep Projective Rotation Estimation through Relative Supervision
Orientation estimation is the core to a variety of vision and robotics tasks such as camera and object pose estimation. Deep learning has offered a way to develop image-based orientation estimators; however, such estimators often require training on a large labeled dataset, which can be time-intensive to collect. In this work, we explore whether self-supervised learning from unlabeled data can be used to alleviate this issue. Specifically, we assume access to estimates of the relative orientation between neighboring poses, such that can be obtained via a local alignment method. While self-supervised learning has been used successfully for translational object keypoints, in this work, we show that naively applying relative supervision to the rotational group $SO(3)$ will often fail to converge due to the non-convexity of the rotational space. To tackle this challenge, we propose a new algorithm for self-supervised orientation estimation which utilizes Modified Rodrigues Parameters to stereographically project the closed manifold of $SO(3)$ to the open manifold of $\mathbb{R}^{3}$, allowing the optimization to be done in an open Euclidean space. We empirically validate the benefits of the proposed algorithm for rotational averaging problem in two settings: (1) direct optimization on rotation parameters, and (2) optimization of parameters of a convolutional neural network that predicts object orientations from images. In both settings, we demonstrate that our proposed algorithm is able to converge to a consistent relative orientation frame much faster than algorithms that purely operate in the $SO(3)$ space. Additional information can be found at https://sites.google.com/view/deep-projective-rotation/home . - 31 October 2022
- Published by Springer Nature
- p. 529-546
Learning Continuous Implicit Representation for Near-Periodic Patterns
- 31 October 2022
- Published by Springer Nature
- p. 529-546
Learning Continuous Implicit Representation for Near-Periodic Patterns
- Open Access
- 22 July 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Graphics
- Vol. 41 (4) , 1-15
Semantically supervised appearance decomposition for virtual staging from a single panorama
We describe a novel approach to decompose a single panorama of an empty indoor environment into four appearance components: specular, direct sunlight, diffuse and diffuse ambient without direct sunlight. Our system is weakly supervised by automatically generated semantic maps (with floor, wall, ceiling, lamp, window and door labels) that have shown success on perspective views and are trained for panoramas using transfer learning without any further annotations. A GAN-based approach supervised by coarse information obtained from the semantic map extracts specular reflection and direct sunlight regions on the floor and walls. These lighting effects are removed via a similar GAN-based approach and a semantic-aware inpainting step. The appearance decomposition enables multiple applications including sun direction estimation, virtual furniture insertion, floor material replacement, and sun direction change, providing an effective tool for virtual home staging. We demonstrate the effectiveness of our approach on a large and recently released dataset of panoramas of empty homes.
- 22 July 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Graphics
- Vol. 41 (4) , 1-15
Open AccessSemantically supervised appearance decomposition for virtual staging from a single panorama
We describe a novel approach to decompose a single panorama of an empty indoor environment into four appearance components: specular, direct sunlight, diffuse and diffuse ambient without direct sunlight. Our system is weakly supervised by automatically generated semantic maps (with floor, wall, ceiling, lamp, window and door labels) that have shown success on perspective views and are trained for panoramas using transfer learning without any further annotations. A GAN-based approach supervised by coarse information obtained from the semantic map extracts specular reflection and direct sunlight regions on the floor and walls. These lighting effects are removed via a similar GAN-based approach and a semantic-aware inpainting step. The appearance decomposition enables multiple applications including sun direction estimation, virtual furniture insertion, floor material replacement, and sun direction change, providing an effective tool for virtual home staging. We demonstrate the effectiveness of our approach on a large and recently released dataset of panoramas of empty homes.
- Open Access
Explainable Semantic Mapping for First Responders
One of the key challenges in the semantic mapping problem in postdisaster environments is how to analyze a large amount of data efficiently with minimal supervision. To address this challenge, we propose a deep learning-based semantic mapping tool consisting of three main ideas. First, we develop a frugal semantic segmentation algorithm that uses only a small amount of labeled data. Next, we investigate on the problem of learning to detect a new class of object using just a few training examples. Finally, we develop an explainable cost map learning algorithm that can be quickly trained to generate traversability cost maps using only raw sensor data such as aerial-view imagery. This paper presents an overview of the proposed idea and the lessons learned.
Open AccessExplainable Semantic Mapping for First Responders
One of the key challenges in the semantic mapping problem in postdisaster environments is how to analyze a large amount of data efficiently with minimal supervision. To address this challenge, we propose a deep learning-based semantic mapping tool consisting of three main ideas. First, we develop a frugal semantic segmentation algorithm that uses only a small amount of labeled data. Next, we investigate on the problem of learning to detect a new class of object using just a few training examples. Finally, we develop an explainable cost map learning algorithm that can be quickly trained to generate traversability cost maps using only raw sensor data such as aerial-view imagery. This paper presents an overview of the proposed idea and the lessons learned.
- Open Access
Edge-Direct Visual Odometry
In this paper we propose an edge-direct visual odometry algorithm that efficiently utilizes edge pixels to find the relative pose that minimizes the photometric error between images. Prior work on exploiting edge pixels instead treats edges as features and employ various techniques to match edge lines or pixels, which adds unnecessary complexity. Direct methods typically operate on all pixel intensities, which proves to be highly redundant. In contrast our method builds on direct visual odometry methods naturally with minimal added computation. It is not only more efficient than direct dense methods since we iterate with a fraction of the pixels, but also more accurate. We achieve high accuracy and efficiency by extracting edges from only one image, and utilize robust Gauss-Newton to minimize the photometric error of these edge pixels. This simultaneously finds the edge pixels in the reference image, as well as the relative camera pose that minimizes the photometric error. We test various edge detectors, including learned edges, and determine that the optimal edge detector for this method is the Canny edge detection algorithm using automatic thresholding. We highlight key differences between our edge direct method and direct dense methods, in particular how higher levels of image pyramids can lead to significant aliasing effects and result in incorrect solution convergence. We show experimentally that reducing the photometric error of edge pixels also reduces the photometric error of all pixels, and we show through an ablation study the increase in accuracy obtained by optimizing edge pixels only. We evaluate our method on the RGB-D TUM benchmark on which we achieve state-of-the-art performance.
Open AccessEdge-Direct Visual Odometry
In this paper we propose an edge-direct visual odometry algorithm that efficiently utilizes edge pixels to find the relative pose that minimizes the photometric error between images. Prior work on exploiting edge pixels instead treats edges as features and employ various techniques to match edge lines or pixels, which adds unnecessary complexity. Direct methods typically operate on all pixel intensities, which proves to be highly redundant. In contrast our method builds on direct visual odometry methods naturally with minimal added computation. It is not only more efficient than direct dense methods since we iterate with a fraction of the pixels, but also more accurate. We achieve high accuracy and efficiency by extracting edges from only one image, and utilize robust Gauss-Newton to minimize the photometric error of these edge pixels. This simultaneously finds the edge pixels in the reference image, as well as the relative camera pose that minimizes the photometric error. We test various edge detectors, including learned edges, and determine that the optimal edge detector for this method is the Canny edge detection algorithm using automatic thresholding. We highlight key differences between our edge direct method and direct dense methods, in particular how higher levels of image pyramids can lead to significant aliasing effects and result in incorrect solution convergence. We show experimentally that reducing the photometric error of edge pixels also reduces the photometric error of all pixels, and we show through an ablation study the increase in accuracy obtained by optimizing edge pixels only. We evaluate our method on the RGB-D TUM benchmark on which we achieve state-of-the-art performance.
- Open Access
Learning Unsupervised Multi-View Stereopsis via Robust Photometric Consistency
We present a learning based approach for multi-view stereopsis (MVS). While current deep MVS methods achieve impressive results, they crucially rely on ground-truth 3D training data, and acquisition of such precise 3D geometry for supervision is a major hurdle. Our framework instead leverages photometric consistency between multiple views as supervisory signal for learning depth prediction in a wide baseline MVS setup. However, naively applying photo consistency constraints is undesirable due to occlusion and lighting changes across views. To overcome this, we propose a robust loss formulation that: a) enforces first order consistency and b) for each point, selectively enforces consistency with some views, thus implicitly handling occlusions. We demonstrate our ability to learn MVS without 3D supervision using a real dataset, and show that each component of our proposed robust loss results in a significant improvement. We qualitatively observe that our reconstructions are often more complete than the acquired ground truth, further showing the merits of this approach. Lastly, our learned model generalizes to novel settings, and our approach allows adaptation of existing CNNs to datasets without ground-truth 3D by unsupervised finetuning. Project webpage: https://tejaskhot.github.io/unsup_mvs
Open AccessLearning Unsupervised Multi-View Stereopsis via Robust Photometric Consistency
We present a learning based approach for multi-view stereopsis (MVS). While current deep MVS methods achieve impressive results, they crucially rely on ground-truth 3D training data, and acquisition of such precise 3D geometry for supervision is a major hurdle. Our framework instead leverages photometric consistency between multiple views as supervisory signal for learning depth prediction in a wide baseline MVS setup. However, naively applying photo consistency constraints is undesirable due to occlusion and lighting changes across views. To overcome this, we propose a robust loss formulation that: a) enforces first order consistency and b) for each point, selectively enforces consistency with some views, thus implicitly handling occlusions. We demonstrate our ability to learn MVS without 3D supervision using a real dataset, and show that each component of our proposed robust loss results in a significant improvement. We qualitatively observe that our reconstructions are often more complete than the acquired ground truth, further showing the merits of this approach. Lastly, our learned model generalizes to novel settings, and our approach allows adaptation of existing CNNs to datasets without ground-truth 3D by unsupervised finetuning. Project webpage: https://tejaskhot.github.io/unsup_mvs