Skip to content
View RUV2005's full-sized avatar
3 followers · 3 following

Block or report RUV2005

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this userโ€™s behavior. Learn more about reporting abuse.

Report abuse
RUV2005/README.md

Intro Typing

Email

Language: English | ไธญๆ–‡

๐Ÿง‘โ€๐Ÿ’ผ About Me

  • ๐Ÿ’ป Focused on innovative fusion of mobile and AI technologies
  • ๐Ÿง  Passionate about audio/video, real-time communication, and system performance
  • ๐ŸŒฑ Currently exploring Edge AI, cross-platform development, and WebAssembly

๐Ÿ“‘ Table of Contents

๐Ÿงฐ Technical Toolbox

Core Competencies

Tech Stack

Deep Expertise

Category Technologies
Mobile Native Android NDK โ€ข Core Audio โ€ข ARCore โ€ข Metal API โ€ข Kotlin Multiplatform
Real-time WebRTC โ€ข gRPC โ€ข WebSocket โ€ข Socket.IO โ€ข LiveData
Architecture Clean Architecture โ€ข MVI โ€ข MVVM โ€ข Event Sourcing โ€ข CQRS
Performance Proguard โ€ข R8 โ€ข Hermes โ€ข V8 Optimization โ€ข SIMD Instructions

๐Ÿš€ Featured Innovations

๐Ÿ‘๏ธ Intelligent Navigation Assistant

TensorFlow Lite Android API Model Accuracy

A real-time spatial awareness and navigation app for the visually impaired, supporting 80+ object detection, sub-400ms end-to-end latency, and 98% device compatibility.

Technical Triumphs:

  • Real-time 80+ object detection with <400ms latency (Optimized EfficientDet-Lite0)
  • 16-direction spatial localization (ยฑ5ยฐ precision)
  • 98% device compatibility (full-series Huawei/Xiaomi support)

Performance Metrics:

โ”œโ”€โ”€ Inference Speed: 350ms (Readmi K80)
โ”œโ”€โ”€ Memory Footprint: 18MB (TensorFlow Lite)
โ”œโ”€โ”€ Power Consumption: 2.3mA avg (Light-sensing mode)
โ””โ”€โ”€ Frame Processing: 21.3fps @ 640x480

System Architecture:

graph TD
    A[CameraX] --> B{Image Preprocessing}
    B --> C[TFLite Inference Engine]
    C --> D[Spatial Coordinate Mapping]
    D --> E[Multi-level Feedback Decision]
    E --> F[TTS Engine]
    E --> G[Vibration Controller]
    E --> H[Ambient Light Adapter]
Loading

๐Ÿ’ก MingMou

TensorFlow Lite Android API

A hybrid online+offline OCR application for low-vision and visually impaired users. Supports classroom PPT/blackboard reading and physical book reading, with seamless switching between the built-in camera and external ESP32-CAM device via AP mode video streaming.

Key Highlights:

  • Both cloud and on-device OCR for robust and private text recognition
  • Specially optimized for classroom: Real-time projection/PPT/blackboard reading
  • ESP32-CAM as an external camera: phone connects via AP mode and streams video for OCR, ideal for reading physical books or documents
  • Seamless hot switching between internal and ESP32-CAM without interruption
  • Fully accessible UI and voice feedback

Core Metrics:

โ”œโ”€โ”€ OCR Latency: <350ms (offline mode)
โ”œโ”€โ”€ ESP32-CAM Switch Time: <1s (hot swap, AP mode)
โ”œโ”€โ”€ Scene Adaptation: Classroom, slides, blackboard, books
โ””โ”€โ”€ Battery Impact: <3mA (typical classroom use)

๐Ÿ–ฅ๏ธ ScreenSaver Motivator

SwiftUI App Store

A dynamic screensaver + motivational quotes engine, integrating calendar, weather, and visual generation for daily inspiration.

Architecture:

graph TD
    A[Context Manager] --> B{Environment Data}
    B --> C[Weather Service]
    B --> D[Calendar Integration]
    A --> E[Content Engine]
    E --> F[Quote Database]
    E --> G[Visual Generator]
    G --> H[SwiftUI Canvas]
    G --> I[Core Animation]
Loading

๐Ÿ“Š Development Analytics

Contribution Graph

Top Langs

๐Ÿงญ Current Exploration Trajectory

  • Edge AI: EfficientDet quantization strategy optimization (98% INT8 accuracy retention)
  • Audio Processing: On-device ML-based noise suppression (TF-Lite)
  • Cross-Platform: Shared KMM audio module for Android/iOS
  • Performance: WebAssembly SIMD optimization for real-time AV processing

๐Ÿ“ฃ Contact

๐Ÿ‡จ๐Ÿ‡ณ ็‚นๅ‡ปๅฑ•ๅผ€ไธญๆ–‡็‰ˆ / Chinese Version

๐Ÿง‘โ€๐Ÿ’ผ ๅ…ณไบŽๆˆ‘

  • ๐Ÿ’ป ไธ“ๆณจ็งปๅŠจ็ซฏไธŽ AI ่žๅˆๅˆ›ๆ–ฐ
  • ๐Ÿง  ็ƒญ่กท้Ÿณ่ง†้ข‘/ๅฎžๆ—ถ้€šไฟก/็ณป็ปŸๆ€ง่ƒฝไผ˜ๅŒ–
  • ๐ŸŒฑ ็›ฎๅ‰ๅ…ณๆณจ Edge AIใ€่ทจ็ซฏๅผ€ๅ‘ใ€WebAssembly

๐Ÿ“‘ ็›ฎๅฝ•

๐Ÿงฐ ๆŠ€ๆœฏๆ ˆ

ๆ ธๅฟƒ่ƒฝๅŠ›

Tech Stack

ๆทฑๅบฆไธ“้•ฟ

ๅˆ†็ฑป ๆŠ€ๆœฏๆ ˆ
็งปๅŠจๅŽŸ็”Ÿ Android NDK โ€ข Core Audio โ€ข ARCore โ€ข Metal API โ€ข Kotlin Multiplatform
ๅฎžๆ—ถ้€šไฟก WebRTC โ€ข gRPC โ€ข WebSocket โ€ข Socket.IO โ€ข LiveData
ๆžถๆž„่ฎพ่ฎก Clean Architecture โ€ข MVI โ€ข MVVM โ€ข Event Sourcing โ€ข CQRS
ๆ€ง่ƒฝไผ˜ๅŒ– Proguard โ€ข R8 โ€ข Hermes โ€ข V8 ไผ˜ๅŒ– โ€ข SIMD ๆŒ‡ไปค้›†

๐Ÿš€ ไปฃ่กจๆ€ง้กน็›ฎ

๐Ÿ‘๏ธ ๆ™บ่ƒฝๅฏผ่ˆชๅŠฉๆ‰‹

TensorFlow Lite Android API Model Accuracy

ไธบ่ง†้šœไบบๅฃซ่ฎพ่ฎก็š„ๅฎžๆ—ถ็ฉบ้—ดๆ„Ÿ็ŸฅไธŽๅฏผ่ˆชๅบ”็”จ๏ผŒๆ”ฏๆŒ 80+ ็›ฎๆ ‡ๆฃ€ๆต‹๏ผŒ็ซฏๅˆฐ็ซฏๅปถ่ฟŸไฝŽไบŽ 400ms๏ผŒๅ…ผๅฎน 98% ไธปๆตๅฎ‰ๅ“่ฎพๅค‡ใ€‚

ๆŠ€ๆœฏไบฎ็‚น:

  • ๅฎžๆ—ถ 80+ ็›ฎๆ ‡ๆฃ€ๆต‹๏ผŒๅปถ่ฟŸ <400ms๏ผˆ้ซ˜ๆ•ˆ EfficientDet-Lite0 ไผ˜ๅŒ–๏ผ‰
  • ้‡‡็”จๅ…ญๅ‘๏ผˆ2ร—3็Ÿฉ้˜ต๏ผ‰็ฉบ้—ดๅฎšไฝๆ–นๅผ๏ผŒๅฎšไฝ็ฒพๅบฆ้ซ˜่พพยฑ5ยฐ๏ผŒ็ฌฆๅˆไบบไฝ“็›ด่ง‚ๆ„Ÿ็Ÿฅไน ๆƒฏ๏ผŒๆ›ดๆ˜“ไบŽ็†่งฃๅ’Œไฝฟ็”จ
  • ่ฎพๅค‡ๅ…ผๅฎน็Ž‡ 98%๏ผˆๅ…จ็ณปๅŽไธบ/ๅฐ็ฑณ้€‚้…๏ผ‰

ๆ€ง่ƒฝๅ‚ๆ•ฐ:

โ”œโ”€โ”€ ๆŽจ็†้€Ÿๅบฆ: 350ms (Readmi K80)
โ”œโ”€โ”€ ๅ†…ๅญ˜ๅ ็”จ: 18MB (TensorFlow Lite)
โ”œโ”€โ”€ ๅŠŸ่€—: 2.3mA ๅนณๅ‡ (ๅ…‰ๆ„Ÿๆจกๅผ)
โ””โ”€โ”€ ๅธงๅค„็†: 21.3fps @ 640x480

็ณป็ปŸๆžถๆž„:

graph TD
    A[CameraX] --> B{ๅ›พๅƒ้ข„ๅค„็†}
    B --> C[TFLite ๆŽจ็†ๅผ•ๆ“Ž]
    C --> D[็ฉบ้—ดๅๆ ‡ๆ˜ ๅฐ„]
    D --> E[ๅคš็บงๅ้ฆˆๅ†ณ็ญ–]
    E --> F[TTS ๅผ•ๆ“Ž]
    E --> G[ๆŒฏๅŠจๆŽงๅˆถๅ™จ]
    E --> H[็Žฏๅขƒๅ…‰้€‚้…ๅ™จ]
Loading

๐Ÿ’ก ๆ˜Ž็œธ (MingMou)

ML Kit Android API

ไธ€ๆฌพ่žๅˆๅœจ็บฟ+็ฆป็บฟ OCR ็š„ๅบ”็”จ๏ผŒไธ“ไธบไฝŽ่ง†ๅŠ›ๅ’Œ่ง†ๅŠ›้šœ็ขไบบๅฃซ่ฎพ่ฎกใ€‚ๆ”ฏๆŒ่ฏพๅ ‚ PPT/้ป‘ๆฟๅ’Œๅฎžไฝ“ไนฆ็š„้˜…่ฏปๅœบๆ™ฏ๏ผŒ้€š่ฟ‡ AP ๆจกๅผ่ฟžๆŽฅ ESP32-CAM ๅฎž็Žฐๅค–็ฝฎๆ‘„ๅƒๅคด่ง†้ข‘ๆต๏ผŒๅ†…็ฝฎ/ๅค–็ฝฎๆ‘„ๅƒๅคดๆ— ็ผ็ƒญๅˆ‡ๆข๏ผŒๆž่‡ดๆ— ้šœ็ขไฝ“้ชŒใ€‚

ไธป่ฆ็‰นๆ€ง:

  • ไบ‘็ซฏ+ๆœฌๅœฐ OCR ๆททๅˆ๏ผŒไฟ้šœ็ฆป็บฟๅŠ้š็ง้œ€ๆฑ‚
  • ้’ˆๅฏน่ฏพๅ ‚็Žฏๅขƒไผ˜ๅŒ–๏ผšPPTใ€้ป‘ๆฟใ€่ฎฒไน‰ๅฎžๆ—ถๆœ—่ฏป
  • ๆ”ฏๆŒๅฎžไฝ“ไนฆใ€ๆ–‡ๆกฃ็ญ‰่ฟ‘่ท็ฆป้˜…่ฏป
  • ๅˆ›ๆ–ฐ ESP32-CAM ๅค–็ฝฎๆ‘„ๅƒๅคดๆŽฅๅ…ฅ๏ผŒๆ‰‹ๆœบ็›ด่ฟž AP๏ผŒไฝŽๅปถ่ฟŸ่ง†้ข‘ๆต OCR
  • ๅ†…็ฝฎ/ESP32-CAM ๆ‘„ๅƒๅคดไธ€้”ฎๆ— ็ผๅˆ‡ๆข
  • ๆ— ้šœ็ขๅ‹ๅฅฝ UI ไธŽ่ฏญ้Ÿณๅ้ฆˆ

ๆ ธๅฟƒๅ‚ๆ•ฐ:

โ”œโ”€โ”€ OCR ๅปถ่ฟŸ: <350ms๏ผˆ็ฆป็บฟๆจกๅผ๏ผ‰
โ”œโ”€โ”€ ESP32-CAM ๅˆ‡ๆข: <1s๏ผˆ็ƒญๅˆ‡ๆข๏ผ‰
โ”œโ”€โ”€ ๅœบๆ™ฏ้€‚้…: ๆ•™ๅฎคใ€ๆŠ•ๅฝฑใ€้ป‘ๆฟใ€ๅฎžไฝ“ไนฆ
โ””โ”€โ”€ ็”ตๆฑ ๅฝฑๅ“: <3mA๏ผˆๅ…ธๅž‹่ฏพๅ ‚ไฝฟ็”จ๏ผ‰

๐Ÿ–ฅ๏ธ ๅŠฑๅฟ—ๅฑไฟ

SwiftUI App Store

ๅŠจๆ€ๅฑไฟ + ๆฟ€ๅŠฑ่ฏญๅฝ•ๅผ•ๆ“Ž๏ผŒ็ป“ๅˆๆ—ฅๅŽ†ใ€ๅคฉๆฐ”ไธŽ่ง†่ง‰็”Ÿๆˆ๏ผŒๆๅ‡ๆ—ฅๅธธไธ“ๆณจๅŠ›ใ€‚

ๆžถๆž„ๅ›พ:

graph TD
    A[Context Manager] --> B{็Žฏๅขƒๆ•ฐๆฎ}
    B --> C[ๅคฉๆฐ”ๆœๅŠก]
    B --> D[ๆ—ฅๅŽ†้›†ๆˆ]
    A --> E[ๅ†…ๅฎนๅผ•ๆ“Ž]
    E --> F[่ฏญๅฝ•ๆ•ฐๆฎๅบ“]
    E --> G[่ง†่ง‰็”Ÿๆˆๅ™จ]
    G --> H[SwiftUI ็”ปๅธƒ]
    G --> I[Core Animation]
Loading

๐Ÿ“Š ๅผ€ๅ‘ๆ•ฐๆฎๅˆ†ๆž

Contribution Graph

Top Langs

๐Ÿงญ ่ฟ‘ๆœŸๆŽข็ดขๆ–นๅ‘

  • Edge AI: EfficientDet ้‡ๅŒ–็ญ–็•ฅไผ˜ๅŒ–๏ผˆINT8 ็ฒพๅบฆไฟ็•™ 98%๏ผ‰
  • ้Ÿณ้ข‘ๅค„็†: ๅŸบไบŽ TF-Lite ็š„็ซฏไพง้™ๅ™ช
  • ่ทจๅนณๅฐ: Android/iOS ๅ…ฑ็”จ KMM ้Ÿณ้ข‘ๆจกๅ—
  • ๆ€ง่ƒฝไผ˜ๅŒ–: WebAssembly SIMD ๅฎžๆ—ถ้Ÿณ่ง†้ข‘ๅŠ ้€Ÿ

๐Ÿ“ฃ ่”็ณปๆ–นๅผ

Code Time

Profile Views

๐Ÿฑ My GitHub Data

๐Ÿ“ฆ 211.5 kB Used in GitHub's Storage

๐Ÿ† 103 Contributions in the Year 2025

๐Ÿšซ Not Opted to Hire

๐Ÿ“œ 14 Public Repositories

๐Ÿ”‘ 2 Private Repositories

I'm an Early ๐Ÿค

๐ŸŒž Morning                62 commits          โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   29.81 % 
๐ŸŒ† Daytime                84 commits          โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   40.38 % 
๐ŸŒƒ Evening                32 commits          โ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   15.38 % 
๐ŸŒ™ Night                  30 commits          โ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   14.42 %

๐Ÿ“… I'm Most Productive on Tuesday

Monday                   45 commits          โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   21.63 % 
Tuesday                  67 commits          โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   32.21 % 
Wednesday                12 commits          โ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   05.77 % 
Thursday                 14 commits          โ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   06.73 % 
Friday                   38 commits          โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   18.27 % 
Saturday                 14 commits          โ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   06.73 % 
Sunday                   18 commits          โ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   08.65 %

๐Ÿ“Š This Week I Spent My Time On

๐Ÿ•‘๏ธŽ Time Zone: Asia/Shanghai

๐Ÿ’ฌ Programming Languages: 
No Activity Tracked This Week

๐Ÿ”ฅ Editors: 
No Activity Tracked This Week

๐Ÿฑโ€๐Ÿ’ป Projects: 
No Activity Tracked This Week

๐Ÿ’ป Operating System: 
No Activity Tracked This Week

I Mostly Code in Kotlin

Kotlin                   5 repos             โ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   41.67 % 
Dart                     1 repo              โ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   08.33 % 
Swift                    1 repo              โ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   08.33 % 
TypeScript               1 repo              โ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   08.33 % 
Shell                    1 repo              โ–ˆโ–ˆโ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘โ–‘   08.33 %

Timeline

Lines of Code chart

Last Updated on 28/10/2025 18:52:29 UTC

Popular repositories Loading

  1. guide guide Public

    Edge device object detection with Kotlin & TensorFlow Lite Mobile & edge AI object detection using Kotlin

    Kotlin 2 1

  2. mingmou mingmou Public

    Kotlin 1

  3. screenproduct screenproduct Public

    C#

  4. network_authentication network_authentication Public

    Python

  5. loud-speaker-android loud-speaker-android Public

    Kotlin 1

  6. nuxt-app nuxt-app Public

    TypeScript