Shubham Jain

Visual navigation has emerged as a practical alternative to traditional robotic navigation pipelines that rely on detailed mapping and path planning. However, constructing and maintaining 3D maps is often computationally expensive and memory-intensive. We address the problem of visual navigation when exploration videos of a large environment are available. The videos serve as a visual reference, allowing a robot to retrace the explored trajectories without relying on metric maps. Our proposed method, YOPO-Nav (You Only Pass Once), encodes an environment into a compact spatial representation composed of interconnected local 3D Gaussian Splatting (3DGS) models. During navigation, the framework aligns the robot's current visual observation with this representation and predicts actions that guide it back toward the demonstrated trajectory. YOPO-Nav employs a hierarchical design: a visual place recognition (VPR) module provides coarse localization, while the local 3DGS models refine the goal and intermediate poses to generate control actions. To evaluate our approach, we introduce the YOPO-Campus dataset, comprising 4 hours of egocentric video and robot controller inputs from over 6 km of human-teleoperated robot trajectories. We benchmark recent visual navigation methods on trajectories from YOPO-Campus using a Clearpath Jackal robot. Experimental results show YOPO-Nav provides excellent performance in image-goal navigation for real-world scenes on a physical robot. The dataset and code will be made publicly available for visual navigation and scene representation research.

Model merging has emerged as a promising solution to accommodate multiple large models within constrained memory budgets. We present StatsMerging, a novel lightweight learning-based model merging method guided by weight distribution statistics without requiring ground truth labels or test samples. StatsMerging offers three key advantages: (1) It uniquely leverages singular values from singular value decomposition (SVD) to capture task-specific weight distributions, serving as a proxy for task importance to guide task coefficient prediction; (2) It employs a lightweight learner StatsMergeLearner to model the weight distributions of task-specific pre-trained models, improving generalization and enhancing adaptation to unseen samples; (3) It introduces Task-Specific Teacher Distillation for merging vision models with heterogeneous architectures, a merging learning paradigm that avoids costly ground-truth labels by task-specific teacher distillation. Notably, we present two types of knowledge distillation, (a) distilling knowledge from task-specific models to StatsMergeLearner; and (b) distilling knowledge from models with heterogeneous architectures prior to merging. Extensive experiments across eight tasks demonstrate the effectiveness of StatsMerging. Our results show that StatsMerging outperforms state-of-the-art techniques in terms of overall accuracy, generalization to unseen tasks, and robustness to image quality variations.

In this work, we present RadGazeGen, a novel framework for integrating experts' eye gaze patterns and radiomic feature maps as controls to text-to-image diffusion models for high fidelity medical image generation. Despite the recent success of text-to-image diffusion models, text descriptions are often found to be inadequate and fail to convey detailed disease-specific information to these models to generate clinically accurate images. The anatomy, disease texture patterns, and location of the disease are extremely important to generate realistic images; moreover the fidelity of image generation can have significant implications in downstream tasks involving disease diagnosis or treatment repose assessment. Hence, there is a growing need to carefully define the controls used in diffusion models for medical image generation. Eye gaze patterns of radiologists are important visuo-cognitive information, indicative of subtle disease patterns and spatial location. Radiomic features further provide important subvisual cues regarding disease phenotype. In this work, we propose to use these gaze patterns in combination with standard radiomics descriptors, as controls, to generate anatomically correct and disease-aware medical images. RadGazeGen is evaluated for image generation quality and diversity on the REFLACX dataset. To demonstrate clinical applicability, we also show classification performance on the generated images from the CheXpert test set (n=500) and long-tailed learning performance on the MIMIC-CXR-LT test set (n=23550).

Tracking subjects in videos is one of the most widely used functions in camera-based IoT applications such as security surveillance, smart city traffic safety enhancement, vehicle to pedestrian communication and so on. In the computer vision domain, tracking is usually achieved by first detecting subjects with bounding boxes, then associating detected bounding boxes across video frames. For many IoT systems, images captured by cameras are usually sent over the network to be processed at a different site that has more powerful computing resources than edge devices. However, sending entire frames through the network causes significant bandwidth consumption that may exceed the system bandwidth constraints. To tackle this problem, we propose ViFiT, a transformer-based model that reconstructs vision bounding box trajectories from phone data (IMU and Fine Time Measurements). It leverages a transformer ability of better modeling long-term time series data. ViFiT is evaluated on Vi-Fi Dataset, a large-scale multimodal dataset in 5 diverse real world scenes, including indoor and outdoor environments. To fill the gap of proper metrics of jointly capturing the system characteristics of both tracking quality and video bandwidth reduction, we propose a novel evaluation framework dubbed Minimum Required Frames (MRF) and Minimum Required Frames Ratio (MRFR). ViFiT achieves an MRFR of 0.65 that outperforms the state-of-the-art approach for cross-modal reconstruction in LSTM Encoder-Decoder architecture X-Translator of 0.98, resulting in a high frame reduction rate as 97.76%.

We examine how the choice of data-side attributes for two important visual tasks of image classification and object detection can aid in the choice or design of lightweight convolutional neural networks. We show by experimentation how four data attributes - number of classes, object color, image resolution, and object scale affect neural network model size and efficiency. Intra- and inter-class similarity metrics, based on metric learning, are defined to guide the evaluation of these attributes toward achieving lightweight models. Evaluations made using these metrics are shown to require 30x less computation than running full inference tests. We provide, as an example, applying the metrics and methods to choose a lightweight model for a robot path planning application and achieve computation reduction of 66% and accuracy gain of 3.5% over the pre-method model.

Recent statistics reveal an alarming increase in accidents involving pedestrians (especially children) crossing the street. A common philosophy of existing pedestrian detection approaches is that this task should be undertaken by the moving cars themselves. In sharp departure from this philosophy, we propose to enlist the help of cars parked along the sidewalk to detect and protect crossing pedestrians. In support of this goal, we propose ADOPT: a system for Alerting Drivers to Occluded Pedestrian Traffic. ADOPT lays the theoretical foundations of a system that uses parked cars to: (1) detect the presence of a group of crossing pedestrians - a crossing cohort; (2) predict the time the last member of the cohort takes to clear the street; (3) send alert messages to those approaching cars that may reach the crossing area while pedestrians are still in the street; and, (4) show how approaching cars can adjust their speed, given several simultaneous crossing locations. Importantly, in ADOPT all communications occur over very short distances and at very low power. Our extensive simulations using SUMO-generated pedestrian and car traffic have shown the effectiveness of ADOPT in detecting and protecting crossing pedestrians.

Patients receiving mechanical ventilation for coronavirus disease 2019 (COVID-19) related, moderate-to-severe acute respiratory distress syndrome (CARDS) have mortality rates between 76–98%. The objective of this retrospective cohort study was to identify differences in prone ventilation effects on oxygenation, pulmonary infiltrates (as observed on chest X-ray (CXR)), and systemic inflammation in CARDS patients by survivorship and to identify baseline characteristics associated with survival after prone ventilation. The study cohort included 23 patients with moderate-to-severe CARDS who received prone ventilation for ≥16 h/day and was segmented by living status: living (n = 6) and deceased (n = 17). Immediately after prone ventilation, PaO₂/FiO₂ improved by 108% (p < 0.03) for the living and 150% (p < 3 × 10⁻⁴) for the deceased. However, the 48 h change in lung infiltrate severity in gravity-dependent lung zones was significantly better for the living than for the deceased (p < 0.02). In CXRs of the lower lungs before prone ventilation, we observed 5 patients with confluent infiltrates bilaterally, 12 patients with ground-glass opacities (GGOs) bilaterally, and 6 patients with mixed infiltrate patterns; 80% of patients with confluent infiltrates were alive vs. 8% of patients with GGOs. In conclusion, our small study indicates that CXRs may offer clinical utility in selecting patients with moderate-to-severe CARDS who will benefit from prone ventilation. Additionally, our study suggests that lung infiltrate severity may be a better indicator of patient disposition after prone ventilation than PaO₂/FiO₂.