arXiv subjects

Mahya Beheshti

Publications and source records attributed to Mahya Beheshti.

10 recordsLinked to original sources

arXiv2026

Evaluating OCR Performance for Assistive Technology: Effects of Walking Speed, Camera Placement, and Camera Type

Optical character recognition (OCR), a process that converts printed or handwritten text into machine-readable form, is widely used in assistive technology for people with blindness and low vision. Yet most evaluations rely on static datasets that do not reflect the challenges of mobile use. In this study, we systematically evaluated OCR performance under both static and dynamic conditions. Static tests measured detection range across distances of 1-7 meters and viewing angles of 0-75 degrees horizontally. Dynamic tests examined the impact of motion by varying walking speed from slow (0.8 m/s) to very fast (1.8 m/s) and compared three camera mounting positions: head-mounted, shoulder-mounted, and handheld. We evaluated both a smartphone and smart glasses, using the phone's main and ultra-wide cameras. Four OCR engines were benchmarked to assess accuracy at different distances and viewing angles: Google Vision, PaddleOCR 3.0, EasyOCR, and Tesseract. PaddleOCR 3.0 was then used to evaluate OCR performance under dynamic walking conditions. Accuracy was computed at the character-level using the Levenshtein ratio against manually defined ground truth. Results showed that recognition accuracy declined with increased walking speed and wider viewing angles. Google Vision achieved the highest overall accuracy, with PaddleOCR close behind as the strongest open-source alternative. Across devices, the phone's main camera achieved the highest accuracy, and a shoulder-mounted placement yielded the highest average among body positions; however, differences among shoulder, head, and hand were not statistically significant.

cs.CV↗

arXiv2026

A Multimodal Assistive System for Product Localization and Retrieval for People who are Blind or have Low Vision

Shopping is a routine activity for sighted individuals, yet for people who are blind or have low vision (pBLV), locating and retrieving products in physical environments remains a challenge. This paper presents a multimodal wearable assistive system that integrates object detection with vision-language models to support independent product or item retrieval, with the goal of enhancing users'autonomy and sense of agency. The system operates through three phases: product search, which identifies target products using YOLO-World detection combined with embedding similarity and color histogram matching; product navigation, which provides spatialized sonification and VLM-generated verbal descriptions to guide users toward the target; and product correction, which verifies whether the user has reached the correct product and provides corrective feedback when necessary. Technical evaluation demonstrated promising performance across all modules, with product detection achieving near-perfect accuracy at close range and high accuracy when facing shelves within 1.5 m. VLM-based navigation achieved up to 94.4% accuracy, and correction accuracy exceeded 86% under optimal model configurations. These results demonstrate the system's potential to address the last-meter problem in assistive shopping. Future work will focus on user studies with pBLV participants and integration with multi-scale navigation ecosystems.

cs.HC↗

arXiv2025

Selective disruption of reach-related saccade timing following a middle-cerebral artery stroke

Background: Coordinated control of eye and hand movements is critical for nearly all goal-directed actions, underpinning tasks ranging from simple object manipulation to complex tool use. This coordination relies on temporal coupling between reach and saccade. Stroke disrupts this process at several levels. Methods: We conducted a comparative eye-tracking study on patients with history of stroke vs. control performing various tasks. We used the Kinereach motion-tracking system with integrated eye-tracking functionality for data collection. Results: Stroke participants showed a flat distribution of saccade timing, with no clear peak or alignment to the onset of the reach. In terms of spatial performance, reach gain, the primary index of functional accuracy, revealed that stroke participants reaches were significantly hypometric compared to those of controls. In the segmented look-then-reach condition, stroke participants demonstrated a restoration of a unimodal saccade timing distribution, time-locked to the saccade cue. Conclusion: Our findings highlight a dissociation between motor execution and coordination in stroke, emphasizing that intact movement components do not guarantee intact motor integration. By identifying a reversible, cue-sensitive disruption in saccade timing linked specifically to its coordination within a reaching movement, this work points toward new opportunities for rehabilitation aimed not only at motor strength and speed, but at training and re-establishing the natural temporal structure of action.

q-bio.NC↗

arXiv2025

Head and Eye Control in Persons with Low Vision during Urban Navigation

Low vision involves a range of visual impairments that significantly impact daily activities, particularly navigation in urban environments. Individuals with low vision often develop adaptive strategies to compensate for visual deficits, relying on head movements to bring objects into their remaining functional field of vision. Research suggests that they focus on road surface markings and building edges to aid in wayfinding and collision avoidance. However, urban navigation presents additional challenges, as obstacles, moving hazards, and tripping dangers may enter their visual loss field, increasing the risk of injury. Traditional eye movement studies are typically conducted in controlled laboratory settings with fixed head positions, limiting the understanding of head-eye coordination in real-world environments. To bridge this gap, we designed a naturalistic, "free-head" experiment using eye-tracking technology to examine head and eye movement patterns during urban navigation. Participants with low vision were compared to a control cohort without visual impairment to test the hypothesis that eye and head movements become decoupled in visually impaired individuals. Findings indicate that individuals with peripheral field loss exhibit significant eye-head decoupling, while those with acuity loss demonstrate more synchronized movements. Results for individuals with central field loss were inconclusive but revealed distinct movement patterns. These insights provide valuable direction for rehabilitation strategies, assistive-mobility technologies, and urban design improvements. By expanding research on eye-head coordination, this study contributes to the development of interventions that enhance safety, mobility, and independence for individuals with low vision in complex urban environments.

q-bio.QM↗

arXiv2025

Residual Gaze Behavior During Navigation in Blindness and Low Vision

Background: Outdoor navigation poses significant challenges for people with blindness or low vision, yet the role of gaze behavior in supporting mobility remains underexplored. Fully sighted individuals typically adopt consistent scanning strategies, whereas those with visual impairments rely on heterogeneous adaptations shaped by residual vision and experience. Methods: We conducted a comparative eye-tracking study of fully sighted, low vision, blind, and fully blind participants navigating outdoor routes. Using a wearable eye tracker, we quantified fixation counts, fixation rate, fixation area, direction, peak fixation location, and walking speed. Results: Walking speed declined systematically with worsening vision. Fixation count increased with greater impairment, reflecting slower travel times and more frequent sampling. Fixation rate rose with worsening vision, though between-group differences were generally not significant between most groups. Fixation spatial coverage decreased along the continuum of vision loss. Fixation patterns were most consistent in the fully sighted group. Peak fixation locations were centered in fully sighted participants but shifted outward and became more variable with impairment. Conclusion: Gaze strategies during navigation form a graded continuum across vision groups, with fully sighted and fully blind participants at opposite poles and low vision and blind groups spanning the middle. Visual acuity alone does not predict functional gaze use, as rehabilitation experience and adaptive strategies strongly shape behavior. These findings highlight the need for personalized rehabilitation and assistive technologies, with residual gaze patterns offering insight into mobility capacity and training opportunities for safer navigation.

q-bio.NC↗

arXiv2024

Multi-faceted Sensory Substitution for Curb Alerting: A Pilot Investigation in Persons with Blindness and Low Vision

Curbs -- the edge of a raised sidewalk at the point where it meets a street -- crucial in urban environments where they help delineate safe pedestrian zones, from dangerous vehicular lanes. However, curbs themselves are significant navigation hazards, particularly for people who are blind or have low vision (pBLV). The challenges faced by pBLV in detecting and properly orientating themselves for these abrupt elevation changes can lead to falls and serious injuries. Despite recent advancements in assistive technologies, the detection and early warning of curbs remains a largely unsolved challenge. This paper aims to tackle this gap by introducing a novel, multi-faceted sensory substitution approach hosted on a smart wearable; the platform leverages an RGB camera and an embedded system to capture and segment curbs in real time and provide early warning and orientation information. The system utilizes YOLO (You Only Look Once) v8 segmentation model, trained on our custom curb dataset for the camera input. The output of the system consists of adaptive auditory beeps, abstract sonification, and speech, conveying information about the relative distance and orientation of curbs. Through human-subjects experimentation, we demonstrate the effectiveness of the system as compared to the white cane. Results show that our system can provide advanced warning through a larger safety window than the cane, while offering nearly identical curb orientation information.

cs.HC↗

arXiv2024

Haptics-based, higher-order Sensory Substitution designed for Object Negotiation in Blindness and Low Vision: Virtual Whiskers

People with blindness and low vision (pBLV) face challenges in navigating. Mobility aids are crucial for enhancing independence and safety. This paper presents an electronic travel aid that leverages a haptic-based, higher-order sensory substitution approach called Virtual Whiskers, designed to help pBLV negotiate obstacles effectively, efficiently, and safely. Virtual Whiskers is equipped with a plurality of modular vibration units that operate independently to deliver haptic feedback to users. Virtual Whiskers features two navigation modes: open path mode and depth mode, each addressing obstacle negotiation from different perspectives. The open path mode detects and delineate a traversable area within an analyzed field of view. Then, it guides the user through to the traversable direction adaptive vibratory feedback. The depth mode assists users in negotiating obstacles by highlighting spatial areas with prominent obstacles via haptic feedback. We recruited 10 participants with blindness or low vision to participate in user testing for Virtual Whiskers. Results show that the device significantly reduces idle periods and decreases the number of cane contacts. Virtual Whiskers is a promising obstacle negotiation strategy that demonstrating great potential to assist with pBLV navigation.

cs.HC↗

arXiv2024

Can Foundation Models Reliably Identify Spatial Hazards? A Case Study on Curb Segmentation

Curbs serve as vital borders that delineate safe pedestrian zones from potential vehicular traffic hazards. Curbs also represent a primary spatial hazard during dynamic navigation with significant stumbling potential. Such vulnerabilities are particularly exacerbated for persons with blindness and low vision (PBLV). Accurate visual-based discrimination of curbs is paramount for assistive technologies that aid PBLV with safe navigation in urban environments. Herein, we investigate the efficacy of curb segmentation for foundation models. We introduce the largest curb segmentation dataset to-date to benchmark leading foundation models. Our results show that state-of-the-art foundation models face significant challenges in curb segmentation. This is due to their high false-positive rates (up to 95%) with poor performance distinguishing curbs from curb-like objects or non-curb areas, such as sidewalks. In addition, the best-performing model averaged a 3.70-second inference time, underscoring problems in providing real-time assistance. In response, we propose solutions including filtered bounding box selections to achieve more accurate curb segmentation. Overall, despite the immediate flexibility of foundation models, their application for practical assistive technology applications still requires refinement. This research highlights the critical need for specialized datasets and tailored model training to address navigation challenges for PBLV and underscores implicit weaknesses in foundation models.

cs.CV↗

arXiv2022

UNav: An Infrastructure-Independent Vision-Based Navigation System for People with Blindness and Low vision

Vision-based localization approaches now underpin newly emerging navigation pipelines for myriad use cases from robotics to assistive technologies. Compared to sensor-based solutions, vision-based localization does not require pre-installed sensor infrastructure, which is costly, time-consuming, and/or often infeasible at scale. Herein, we propose a novel vision-based localization pipeline for a specific use case: navigation support for end-users with blindness and low vision. Given a query image taken by an end-user on a mobile application, the pipeline leverages a visual place recognition (VPR) algorithm to find similar images in a reference image database of the target space. The geolocations of these similar images are utilized in downstream tasks that employ a weighted-average method to estimate the end-user's location and a perspective-n-point (PnP) algorithm to estimate the end-user's direction. Additionally, this system implements Dijkstra's algorithm to calculate a shortest path based on a navigable map that includes trip origin and destination. The topometric map used for localization and navigation is built using a customized graphical user interface that projects a 3D reconstructed sparse map, built from a sequence of images, to the corresponding a priori 2D floor plan. Sequential images used for map construction can be collected in a pre-mapping step or scavenged through public databases/citizen science. The end-to-end system can be installed on any internet-accessible device with a camera that hosts a custom mobile application. For evaluation purposes, mapping and localization were tested in a complex hospital environment. The evaluation results demonstrate that our system can achieve localization with an average error of less than 1 meter without knowledge of the camera's intrinsic parameters, such as focal length.

cs.CV↗

arXiv2022

Network-Aware 5G Edge Computing for Object Detection: Augmenting Wearables to "See" More, Farther and Faster

Advanced wearable devices are increasingly incorporating high-resolution multi-camera systems. As state-of-the-art neural networks for processing the resulting image data are computationally demanding, there has been growing interest in leveraging fifth generation (5G) wireless connectivity and mobile edge computing for offloading this processing to the cloud. To assess this possibility, this paper presents a detailed simulation and evaluation of 5G wireless offloading for object detection within a powerful, new smart wearable called VIS4ION, for the Blind-and-Visually Impaired (BVI). The current VIS4ION system is an instrumented book-bag with high-resolution cameras, vision processing and haptic and audio feedback. The paper considers uploading the camera data to a mobile edge cloud to perform real-time object detection and transmitting the detection results back to the wearable. To determine the video requirements, the paper evaluates the impact of video bit rate and resolution on object detection accuracy and range. A new street scene dataset with labeled objects relevant to BVI navigation is leveraged for analysis. The vision evaluation is combined with a detailed full-stack wireless network simulation to determine the distribution of throughputs and delays with real navigation paths and ray-tracing from new high-resolution 3D models in an urban environment. For comparison, the wireless simulation considers both a standard 4G-Long Term Evolution (LTE) carrier and high-rate 5G millimeter-wave (mmWave) carrier. The work thus provides a thorough and realistic assessment of edge computing with mmWave connectivity in an application with both high bandwidth and low latency requirements.

eess.IV↗