Back
Case Study Dossier2024

SmartAir Mouse Pro: Zero-Latency Touchless OS Control

Overcoming sensor noise & tremor with a 2D Kalman filter on 21 MediaPipe hand landmarks

Role
Computer Vision & AI Engineer
Team & Context
Independent Research & Engineering Project
Key Impact
60 FPS Smooth Desktop Cursor Control
GitHub

1. The Problem

Touchless computer vision mouse systems usually suffer from an unavoidable dilemma: without smoothing, hand tremors cause violent cursor jitter; with traditional moving-average filters, the cursor lags by 150-200ms, making accurate clicking virtually impossible.

2. Hard Engineering Constraints

  • Must run smoothly at 60 FPS on integrated laptop webcams without requiring a dedicated GPU.
  • Total pipeline latency from camera photon to OS cursor dispatch must remain under 30ms.
  • Reliable pinch-to-click detection without false positives caused by accidental finger drift.

3. System Architecture Topology

[Webcam Stream (720p 60fps)] ──> [OpenCV Frame Preprocessor]
                                               │
                                  [MediaPipe 21-pt Hand Mesh]
                                               │
                                   [2D Kalman State Estimator]
                                               │
                         ┌─────────────────────┴─────────────────────┐
                         ▼                                           ▼
             [OS Cursor Dispatch (Win32)]                 [Neon Particle AR HUD]

4. My Key Architectural Contributions

Dynamic 2D Kalman Filter Tuning

Implemented state-space prediction matrices calibrated specifically for human micro-tremors, predicting future coordinates to cancel sensor delay while aggressively dampening static jitter.

Euclidean Pinch Distance Calibration

Developed an adaptive depth-normalized distance metric between index tip (landmark 8) and thumb tip (landmark 4) that dynamically scales based on palm bounding box dimensions.

Kinetic Particle AR HUD & Gradio Sandbox

Designed a transparent overlay that explodes visual particle physics on click events, and deployed a browser-accessible Gradio demo for real-time hyperparameter tuning.

5. Measured Engineering Outcomes

FPS

Locked frame rate on modest CPU hardware (Core i5 integrated graphics)

<ms

Total pipeline latency (frame capture -> landmark -> Kalman -> dispatch)

%

Precision on intentional click gestures vs incidental finger movement

6. Retrospective & Senior Engineering Review

What Worked Well

The Kalman filter completely solved the jitter vs lag problem that plagued simpler moving-average implementations.

What I'd Improve in v2

In a future release, I would replace the Win32 mouse API with cross-platform virtual HID drivers to support Linux and macOS touchpads.

Technology Stack & Tooling

PythonOpenCVMediaPipeKalman FilterGradioWindows OS APIs
Related Work

Explore Related Case Studies

Explore real engineering results across cloud backends, AI agents, and computer vision systems.

Tensor Security Academy — Platform & Automation preview

Tensor Security Academy — Platform & Automation

“Automated booking dispatch via n8n and hardened Next.js ISR architectures delivering 99.9% uptime with zero manual database intervention.”
Next.js App RouterTypeScriptn8n AutomationPostgreSQLDockerVercel
Saobia Islam TinniLead Software Engineer (Independent Maintainer)
Dossier
Live

%

Platform Uptime

Enterprise SLA across global students

0 hrs

Manual Operations

Full n8n webhook booking automation

A+

Security Headers Rating

Mozilla Observatory & SSL Labs

SSRN Transfer — US Fleet & Booking Platform preview

SSRN Transfer — US Fleet & Booking Platform

“Engineered high-throughput responsive vehicle reservation templates with zero-lag debounced fleet filtering and mobile-first booking workflows.”
JavaScriptHTML5CSS3Jinja2PythonResponsive Systems
Saobia Islam TinniFrontend Developer (Project-Based)
Dossier

ms

Query Response

Debounced fleet availability filtering

FPS

Render Performance

CLS < 0.001 mobile zero-lag interaction

%

Cross-Device Reliability

Tested across iOS Safari & Android

Arelix Labs — AI Agency & Engineering Hub preview

Arelix Labs — AI Agency & Engineering Hub

“Architected full-stack Next.js 15 App Router platform with Prisma ORM, Neon Postgres, and Groq LLM pipelines for automated scope generation.”
Next.js App RouterTypeScriptPrisma ORMNeon PostgresGroq AI SDK
Saobia Islam TinniFounding Engineer & CTO
Dossier
Live

ms

Inference Routing Latency

Groq LPU accelerated streaming

x

Scope Generation Speed

Automated proposal-to-deliverable AI

%

Transaction Reliability

Connection pooling & rollback protection

AUST Robotics Club & ARC Event Platform preview

AUST Robotics Club & ARC Event Platform

“Engineered a schema-driven monorepo platform powering nationwide robotics tournaments with atomic Firestore scoring and ImageKit CDN.”
ReactTypeScriptMonorepoFirebase FirestoreImageKit CDN
Saobia Islam TinniDirector of Web & App Development Team
Dossier
Live

%

Code Reusability

Zero-code re-theming per edition

National Championships

Powering ARC 2.0, Robomania & Club

ms

Asset Delivery

ImageKit edge cache verification cards

Tensora — Agentic AI Platform & Chatbot preview

Tensora — Agentic AI Platform & Chatbot

“Multi-model reasoning agent system with token streaming over HTTP/2 SSE and dynamic cost guardrails cutting inference spend by 42%.”
Next.jsTypeScriptGroq AI SDKLangGraphFastAPIPinecone
Saobia Islam TinniLead Software Engineer & System Architect
Dossier
Live

ms

First-Token Latency

Streaming SSE transformer pipeline

%

Cost Optimization

Frontier-to-lightweight model routing

%

Tool Call Precision

Deterministic DB + vector search