Avideh Zakhor

Lucas Jaffe, Avideh Zakhor

In person search, we aim to localize a query person from one scene in other gallery scenes. The cost of this search operation is dependent on the number of gallery scenes, making it beneficial to reduce the pool of likely scenes. We describe and demonstrate the Gallery Filter Network (GFN), a novel module which can efficiently discard gallery scenes from the search process, and benefit scoring for persons detected in remaining scenes. We show that the GFN is robust under a range of different conditions by testing on different retrieval sets, including cross-camera, occluded, and low-resolution scenarios. In addition, we develop the base SeqNeXt person search model, which improves and simplifies the original SeqNet model. We show that the SeqNeXt+GFN combination yields significant performance gains over other state-of-the-art methods on the standard PRW and CUHK-SYSU person search datasets. To aid experimentation for this and other models, we provide standardized tooling for the data processing and evaluation pipeline typically used for person search research.

Monica Tang, Avideh Zakhor

We present an approach for object-level detection and segmentation of target indoor assets in 3D Gaussian Splatting (3DGS) scenes, reconstructed from 360° drone-captured imagery. We introduce a 3D object codebook that jointly leverages mask semantics and spatial information of their corresponding Gaussian primitives to guide multi-view mask association and indoor asset detection. By integrating 2D object detection and segmentation models with semantically and spatially constrained merging procedures, our method aggregates masks from multiple views into coherent 3D object instances. Experiments on two large indoor scenes demonstrate reliable multi-view mask consistency, improving F1 score by 65% over state-of-the-art baselines, and accurate object-level 3D indoor asset detection, achieving an 11% mAP gain over baseline methods.

Yoga Satwik Chappidi, Avideh Zakhor

Autonomous exploration is critical for robot mapping unknown environments. Desirable characteristics of exploration algorithms include compute efficiency and small traversed distance during the exploration process. Motivated by these, we present Omnidirectional Path-efficient Aerial 3D expLoration (OPAL), an exploration framework centered on deliberate 360-degree yaw rotation at ambiguous branch points rather than compute-heavy global tour planning. We devise multiple variants of OPAL to determine the frontier-selection strategy once the yaw pan is completed. One variant is model-free, while others use large language models (LLMs) or vision-language models (VLMs). We characterize the performance of these variants while varying the vicinity search radius to include frontiers in the selection process. Through simulations we find that although the time-consuming in-place yaw rotation increases total exploration time relative to more computationally complex baselines such as EDEN and FALCON, OPAL is computationally simpler and achieves shorter travel distances and higher coverage-versus-distance area under the curve. We also show that adjusting the frontier-selection search radius enables a tradeoff between travel distance and total exploration time. We verify our results on a Modal AI drone in two indoor environments by comparing OPAL against FALCON, and find that the traveled distance for a variant of OPAL to be as much as 25% lower than FALCON.

Shiladitya Dutta, Aayush Gupta, Varun Saran et al.

Although quadcopters boast impressive traversal capabilities enabled by their omnidirectional maneuverability, the need for continuous pilot control in complex environments impedes their application in GNSS and telemetry-denied scenarios. To this end, we propose a novel sensorimotor policy that uses stereo-vision depth and visual-inertial odometry (VIO) to autonomously navigate through obstacles in an unknown environment to reach a goal point. The policy is comprised of a pre-trained autoencoder as the perception head followed by a planning and control LSTM network which outputs velocity commands that can be followed by an off-the-shelf commercial drone. We leverage reinforcement and privileged learning paradigms to train the policy in simulation through a two-stage process: 1) initial training with optimal trajectories generated by a global motion planner acting as a supervisory backbone, 2) further fine-tuning in a curriculum environment. To bridge the sim-to-real gap, we employ domain randomization and reward shaping to create a policy that is both robust to noise and domain shift. In outdoor experiments, our approach achieves successful zero-shot transfer to both obstacle environments and a drone platform that were never encountered during training.

Aarush Aitha, Avideh Zakhor

Autonomous robotic exploration of unknown and hazardous environments, a long-standing challenge, can be significantly improved by leveraging the advanced reasoning of Vision-Language Models (VLMs). We introduce a novel exploration pipeline where a VLM performs high-level strategic decision-making, guiding a conventional low-level robotics control stack. At decision points, the robot generates a multimodal prompt with its current map and visual imagery of potential paths, or frontiers. The VLM analyzes this prompt to select the most promising frontier, replacing simple geometric heuristics with contextual spatial reasoning. This approach, validated in simulation across six indoor environments, improves map coverage by up to 24\% over existing methods. Our pipeline is lightweight, training-free, and easily transferable to any robot with standard sensors and an internet connection.

Nitin Vegesna, Avideh Zakhor

We present Semantic-Aware Guided Exploration, SAGE, a system for open-vocabulary exploration in unknown 3D indoor environments that preserves coverage-oriented behavior while allowing semantic cues to reprioritize frontier selection. Building on the FALCON volumetric explorer, SAGE integrates Contrastive Language-Image Pre-training (CLIP) via four key components: object-centric embedding storage, a temporal cache that projects recent observations onto the free-unknown boundary, object frontiers for high-similarity detections, and a unified semantic-geometric planning cost. This cost function bounds semantic reweighting influence, ensuring frontiers are prioritized without sacrificing total coverage. In Matterport3D-based simulations, SAGE outperforms FALCON and a semantic-only ablation in object discovery across map-query pairs. Compared to Finding Things in the Unknown (FTU), SAGE completes exploration 9.0 to 25.9 times faster across the nine shared map-query pairs, achieving a mean speedup of 13.7. Furthermore, SAGE achieves substantially higher volumetric throughput than FTU. Finally, we deploy SAGE in five real-world flights in two environments on a Modal AI Starling 2 quadrotor with onboard sensing and planning, and offboard CLIP inference. Comparing SAGE and FALCON, we find that while FALCON results in faster exploration and shorter mapping trajectories, SAGE outperforms FALCON in terms of object discovery.

Jerome Quenum, Kehan Wang, Avideh Zakhor

Object detection in Ultra High-Resolution (UHR) images has long been a challenging problem in computer vision due to the varying scales of the targeted objects. When it comes to barcode detection, resizing UHR input images to smaller sizes often leads to the loss of pertinent information, while processing them directly is highly inefficient and computationally expensive. In this paper, we propose using semantic segmentation to achieve a fast and accurate detection of barcodes of various scales in UHR images. Our pipeline involves a modified Region Proposal Network (RPN) on images of size greater than 10k$\times$10k and a newly proposed Y-Net segmentation network, followed by a post-processing workflow for fitting a bounding box around each segmented barcode mask. The end-to-end system has a latency of 16 milliseconds, which is $2.5\times$ faster than YOLOv4 and $5.9\times$ faster than Mask R-CNN. In terms of accuracy, our method outperforms YOLOv4 and Mask R-CNN by a $mAP$ of 5.5% and 47.1% respectively, on a synthetic dataset. We have made available the generated synthetic barcode dataset and its code at http://www.github.com/viplabB/SBD/.

Lucas Jaffe, Avideh Zakhor

In person search, we detect and rank matches to a query person image within a set of gallery scenes. Most person search models make use of a feature extraction backbone, followed by separate heads for detection and re-identification. While pre-training methods for vision backbones are well-established, pre-training additional modules for the person search task has not been previously examined. In this work, we present the first framework for end-to-end person search pre-training. Our framework splits person search into object-centric and query-centric methodologies, and we show that the query-centric framing is robust to label noise, and trainable using only weakly-labeled person bounding boxes. Further, we provide a novel model dubbed Swap Path Net (SPNet) which implements both query-centric and object-centric training objectives, and can swap between the two while using the same weights. Using SPNet, we show that query-centric pre-training, followed by object-centric fine-tuning, achieves state-of-the-art results on the standard PRW and CUHK-SYSU person search benchmarks, with 96.4% mAP on CUHK-SYSU and 61.2% mAP on PRW. In addition, we show that our method is more effective, efficient, and robust for person search pre-training than recent backbone-only pre-training alternatives.

Yuanbo Chen, Chengyu Zhang, Jason Wang et al.

Scene reconstruction and novel-view synthesis for large, complex, multi-story, indoor scenes is a challenging and time-consuming task. Prior methods have utilized drones for data capture and radiance fields for scene reconstruction, both of which present certain challenges. First, in order to capture diverse viewpoints with the drone's front-facing camera, some approaches fly the drone in an unstable zig-zag fashion, which hinders drone-piloting and generates motion blur in the captured data. Secondly, most radiance field methods do not easily scale to arbitrarily large number of images. This paper proposes an efficient and scalable pipeline for indoor novel-view synthesis from drone-captured 360 videos using 3D Gaussian Splatting. 360 cameras capture a wide set of viewpoints, allowing for comprehensive scene capture under a simple straightforward drone trajectory. To scale our method to large scenes, we devise a divide-and-conquer strategy to automatically split the scene into smaller blocks that can be reconstructed individually and in parallel. We also propose a coarse-to-fine alignment strategy to seamlessly match these blocks together to compose the entire scene. Our experiments demonstrate marked improvement in both reconstruction quality, i.e. PSNR and SSIM, and computation time compared to prior approaches.

Kehan Wang, David Chan, Seth Z. Zhao et al.

With the growing adoption of short-form video by social media platforms, reducing the spread of misinformation through video posts has become a critical challenge for social media providers. In this paper, we develop methods to detect misinformation in social media posts, exploiting modalities such as video and text. Due to the lack of large-scale public data for misinformation detection in multi-modal datasets, we collect 160,000 video posts from Twitter, and leverage self-supervised learning to learn expressive representations of joint visual and textual data. In this work, we propose two new methods for detecting semantic inconsistencies within short-form social media video posts, based on contrastive learning and masked language modeling. We demonstrate that our new approaches outperform current state-of-the-art methods on both artificial data generated by random-swapping of positive samples and in the wild on a new manually-labeled test set for semantic misinformation.

Maxime Kawawa-Beaudan, Ryan Roggenkemper, Avideh Zakhor

Learned image compression methods generally optimize a rate-distortion loss, trading off improvements in visual distortion for added bitrate. Increasingly, however, compressed imagery is used as an input to deep learning networks for various tasks such as classification, object detection, and superresolution. We propose a recognition-aware learned compression method, which optimizes a rate-distortion loss alongside a task-specific loss, jointly learning compression and recognition networks. We augment a hierarchical autoencoder-based compression network with an EfficientNet recognition model and use two hyperparameters to trade off between distortion, bitrate, and recognition performance. We characterize the classification accuracy of our proposed method as a function of bitrate and find that for low bitrates our method achieves as much as 26% higher recognition accuracy at equivalent bitrates compared to traditional methods such as Better Portable Graphics (BPG).

Lizhi Yang, Ruhang Ma, Avideh Zakhor

Object detection using aerial drone imagery has received a great deal of attention in recent years. While visible light images are adequate for detecting objects in most scenarios, thermal cameras can extend the capabilities of object detection to night-time or occluded objects. As such, RGB and Infrared (IR) fusion methods for object detection are useful and important. One of the biggest challenges in applying deep learning methods to RGB/IR object detection is the lack of available training data for drone IR imagery, especially at night. In this paper, we develop several strategies for creating synthetic IR images using the AIRSim simulation engine and CycleGAN. Furthermore, we utilize an illumination-aware fusion framework to fuse RGB and IR images for object detection on the ground. We characterize and test our methods for both simulated and actual data. Our solution is implemented on an NVIDIA Jetson Xavier running on an actual drone, requiring about 28 milliseconds of processing per RGB/IR image pair.

Zach Van Hyfte, Avideh Zakhor

Smartphone apps for exposure notification and contact tracing have been shown to be effective in controlling the COVID-19 pandemic. However, Bluetooth Low Energy tokens similar to those broadcast by existing apps can still be picked up far away from the transmitting device. In this paper, we present a new class of methods for detecting whether or not two Wi-Fi-enabled devices are in immediate physical proximity, i.e. 2 or fewer meters apart, as established by the U.S. Centers for Disease Control and Prevention (CDC). Our goal is to enhance the accuracy of smartphone-based exposure notification and contact tracing systems. We present a set of binary machine learning classifiers that take as input pairs of Wi-Fi RSSI fingerprints. We empirically verify that a single classifier cannot generalize well to a range of different environments with vastly different numbers of detectable Wi-Fi Access Points (APs). However, specialized classifiers, tailored to situations where the number of detectable APs falls within a certain range, are able to detect immediate physical proximity significantly more accurately. As such, we design three classifiers for situations with low, medium, and high numbers of detectable APs. These classifiers distinguish between pairs of RSSI fingerprints recorded 2 or fewer meters apart and pairs recorded further apart but still in Bluetooth range. We characterize their balanced accuracy for this task to be between 66.8% and 77.8%.

Scott McCrae, Kehan Wang, Avideh Zakhor

As computer-generated content and deepfakes make steady improvements, semantic approaches to multimedia forensics will become more important. In this paper, we introduce a novel classification architecture for identifying semantic inconsistencies between video appearance and text caption in social media news posts. We develop a multi-modal fusion framework to identify mismatches between videos and captions in social media posts by leveraging an ensemble method based on textual analysis of the caption, automatic audio transcription, semantic video analysis, object detection, named entity consistency, and facial verification. To train and test our approach, we curate a new video-based dataset of 4,000 real-world Facebook news posts for analysis. Our multi-modal approach achieves 60.5% classification accuracy on random mismatches between caption and appearance, compared to accuracy below 50% for uni-modal models. Further ablation studies confirm the necessity of fusion across modalities for correctly identifying semantic inconsistencies.

David Deng, Avideh Zakhor

Anticipating the future in a dynamic scene is critical for many fields such as autonomous driving and robotics. In this paper we propose a class of novel neural network architectures to predict future LiDAR frames given previous ones. Since the ground truth in this application is simply the next frame in the sequence, we can train our models in a self-supervised fashion. Our proposed architectures are based on FlowNet3D and Dynamic Graph CNN. We use Chamfer Distance (CD) and Earth Mover's Distance (EMD) as loss functions and evaluation metrics. We train and evaluate our models using the newly released nuScenes dataset, and characterize their performance and complexity with several baselines. Compared to directly using FlowNet3D, our proposed architectures achieve CD and EMD nearly an order of magnitude lower. In addition, we show that our predictions generate reasonable scene flow approximations without using any labelled supervision.

Mohammad Keshavarzi, Oladapo Afolabi, Luisa Caldas et al.

The availability of rich 3D datasets corresponding to the geometrical complexity of the built environments is considered an ongoing challenge for 3D deep learning methodologies. To address this challenge, we introduce GenScan, a generative system that populates synthetic 3D scan datasets in a parametric fashion. The system takes an existing captured 3D scan as an input and outputs alternative variations of the building layout including walls, doors, and furniture with corresponding textures. GenScan is a fully automated system that can also be manually controlled by a user through an assigned user interface. Our proposed system utilizes a combination of a hybrid deep neural network and a parametrizer module to extract and transform elements of a given 3D scan. GenScan takes advantage of style transfer techniques to generate new textures for the generated scenes. We believe our system would facilitate data augmentation to expand the currently limited 3D geometry datasets commonly used in 3D computer vision, generative design, and general 3D deep learning tasks.

Ilya Chugunov, Avideh Zakhor

Static gesture recognition is an effective non-verbal communication channel between a user and their devices; however many modern methods are sensitive to the relative pose of the user's hands with respect to the capture device, as parts of the gesture can become occluded. We present two methodologies for gesture recognition via synchronized recording from two depth cameras to alleviate this occlusion problem. One is a more classic approach using iterative closest point registration to accurately fuse point clouds and a single PointNet architecture for classification, and the other is a dual Point-Net architecture for classification without registration. On a manually collected data-set of 20,100 point clouds we show a 39.2% reduction in misclassification for the fused point cloud method, and 53.4% for the dual PointNet, when compared to a standard single camera pipeline.