Skip to content

Latest commit

 

History

History
248 lines (157 loc) · 6.74 KB

README.md

File metadata and controls

248 lines (157 loc) · 6.74 KB

Google Developer Student Clubs 2024 Solution Challenge


Top 100 !! 🥳🥳


wanna see commits? Go to Part Repo!

iOS

https://github.com/GDSC-PathPal/PathPal-iOS


AI/ML

https://github.com/GDSC-PathPal/PathPal-ML


Backend

https://github.com/GDSC-PathPal/PathPal-Server


표지

‣ Introduction

PathPal is a solution helps visually impaired people walk safely. PathPal provides an accurate route is provided through voice guidance from the starting point to the destination. In addition, it collects visual information through cameras, detects objects and analyzes them using AI/ML to informs users to prevent risk factors through voice and vibration.


Watch the product demo on:

for top10
Youtube


for top100
Youtube

‣ Initial sruvey and Probelm statement research

complaints

overview-h p

needs of assistive device

시장규모
  • Of the 1.1 billion visually impared around the world, 44% have severe disabilities.
  • Additionally, the global assistive device market size exceeded 14 billion$ in 2015. It is expected to reach approximately 25.6 billion$ in 2026.



‣ Mockup

Light Mode(English)

목업 라이트

Dark Mode(Korean)

목업 다크
  1. users can receive voice guidance for the route when they enter thier starting and arraival point.
  2. If user is heading a correct starting direction, navigation will begin with vibration.
  3. results of the camera screen analyzed by stride of the visually impaired person, will be announced through voice and vibration.
  4. In the example screen, you can receive voice guidance such as, “crosswalk in the center and bollard on the right detected”



‣ Key Features

Navigation

  • Current location and heading Management
  • Voice search for starting and destination
  • Setting the correct starting direction using a compass

Vision Assistance

  • Voice guidance on objects and their locations shown on the camera screen
    • Detectable objects : beverage vending machine, bicycle road, braille sign, brailleblock, chair, resting place, lift , crosswalk, trashcan, green traffic
  • Vibration guidance when detecting an accident-causing obstacle
    • Detectable Risks : block kind bad, raised curb, Pillar, bollard, barricade

Globalization




‣ UN SDGs targets

목표 SDGs


‣ Expected Values

기대 효과

Architecture


아키텍쳐

iOS

Guidance

You can seek swift code following steps below.

  1. Go to PathPal directory
  2. Since iOS code of PathPal follow MNVVM patter, you can check View, ViewModel, Model code in directories.
  3. Localized string keys for multilingualn is in en.lproj


AI/ML

Data label collection (before filtering)

ML레이블

Data label collection (after filtering) - actually providing on service

bold - with vibration alert

Pavement status

  • flatness_D, flatness_E
  • paved_state_broken
  • block_kind_bad
  • outcurb_rectangle, outcurb_slide, outcurb_rectengle_broken

Obstacles on sidewalk

  • sidegap_out
  • steepramp

Brailleblock

  • brailleblock_dot, brailleblock_line, brailleblock_dot_broken, brailleblock_line_broken
  • bicycleroad_broken, bicycleroad_normal

Crosswalk

  • planecrosswalk_broken, planecrosswalk_normal

Means of transportation

  • restsapce
  • stair_normal, stair_broken
  • lift

Signs

  • sign_disabled_toilet
  • braille_sign

Infra and Resting places

  • resting_place_roof
  • chair_multi, chair_back, chair_handle
  • beverage_vending_machine
  • trash_can

Backend

server architecture

‣ Getting Started

Download Femunity directly from our GitHub repository. After downloading the app, you can sign up for an account using your Google account or use Guest Mode to explore the app.

Prerequisites

Before you start, make sure you have installed the following on your system:

  • JDK 17

Installation

ML & Server

iOS ML & Server

‣ License

PathPal-Server is licensed under the terms of the MIT license. See License.txt for more information