Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DISHA: Low-Energy Sparse Transformer at Edge for Outdoor Navigation for the Visually Impaired Individuals

Praveen Nagil, Sumit K. Mandal

Abstract

Assistive technology for visually impaired individuals is extremely useful to make them independent of another human being in performing day-to-day chores and instill confidence in them. One of the important aspects of assistive technology is outdoor navigation for visually impaired people. While there exist several techniques for outdoor navigation in the literature, they are mainly limited to obstacle detection. However, navigating a visually impaired person through the sidewalk (while the person is walking outside) is important too. Moreover, the assistive technology should ensure low-energy operation to extend the battery life of the device. Therefore, in this work, we propose an end-to-end technology deployed on an edge device to assist visually impaired people. Specifically, we propose a novel pruning technique for transformer algorithm which detects sidewalk. The pruning technique ensures low latency of execution and low energy consumption when the pruned transformer algorithm is deployed on the edge device. Extensive experimental evaluation shows that our proposed technology provides up to 32.49% improvement in accuracy and 1.4 hours of extension in battery life with respect to a baseline technique.

DISHA: Low-Energy Sparse Transformer at Edge for Outdoor Navigation for the Visually Impaired Individuals

Abstract

Assistive technology for visually impaired individuals is extremely useful to make them independent of another human being in performing day-to-day chores and instill confidence in them. One of the important aspects of assistive technology is outdoor navigation for visually impaired people. While there exist several techniques for outdoor navigation in the literature, they are mainly limited to obstacle detection. However, navigating a visually impaired person through the sidewalk (while the person is walking outside) is important too. Moreover, the assistive technology should ensure low-energy operation to extend the battery life of the device. Therefore, in this work, we propose an end-to-end technology deployed on an edge device to assist visually impaired people. Specifically, we propose a novel pruning technique for transformer algorithm which detects sidewalk. The pruning technique ensures low latency of execution and low energy consumption when the pruned transformer algorithm is deployed on the edge device. Extensive experimental evaluation shows that our proposed technology provides up to 32.49% improvement in accuracy and 1.4 hours of extension in battery life with respect to a baseline technique.
Paper Structure (14 sections, 7 equations, 7 figures, 3 tables)

This paper contains 14 sections, 7 equations, 7 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Overview of the Proposed End-to-end Navigation System
  4. Constructing Sparse Transformer Algorithm
  5. Overview of the Pruning Framework
  6. Computing Pruning Ratio ($p_b$) for each Transformer Block
  7. Pruning each Transformer Block using $p_b$
  8. End-to-end System Implementation
  9. Experimental Evaluation
  10. Experimental Setup
  11. Demonstration of k-score and $p_b$ Calculation
  12. Comparison in Accuracy
  13. Improvement in Execution Latency and Energy Consumption
  14. Conclusion

Figures (7)

  • Figure 1: Overview of the Proposed Outdoor Navigation System for Visually Impaired Individuals -- DISHA. (a) shows real time usage and (b) shows the working principle.
  • Figure 2: A transformer algorithm as seen by our proposed pruning framework.
  • Figure 3: End-to-end implementation of DISHA.
  • Figure 4: Comparison of accuracy with mapillary vistas dataset with input pruning ratio of 0.35.
  • Figure 5: Comparison of accuracy with mapillary vistas dataset with input pruning ratio of 0.40.
  • ...and 2 more figures