AI & Computer Vision

Cloth Measurement - AI Body Sizing & Garment Measurement API

AI body-measurement service that derives garment sizes from photos using MediaPipe pose detection. Operators submit front and side body images (or a real-time capture session), and the service computes key measurements - height, neck, shoulder, chest, waist, hip, arm, inseam and outseam - converting pixel distances to real-world cm via a scale factor.

01 Overview

AI body-measurement service that derives garment sizes from photos using MediaPipe pose detection. Operators submit front and side body images (or a real-time capture session), and the service computes key measurements - height, neck, shoulder, chest, waist, hip, arm, inseam and outseam - converting pixel distances to real-world cm via a scale factor.

Scope of delivery

Built the Cloth Measurement body-sizing service (body_measurements branch). Developed a FastAPI backend with a MediaPipe-based PoseDetector that tracks key body landmarks (nose, shoulders, wrists, ankles) and a MeasurementCalculator that converts pixel landmarks into real-world garment measurements using a height-normalized scale factor, exposing stateful capture-session endpoints for front and side photo / real-time input.

02 Business Challenge

AI body-measurement service that derives garment sizes from photos using MediaPipe pose detection. Operators submit front and side body images (or a real-time capture session), and the service computes key measurements - height, neck, shoulder, chest, waist, hip, arm, inseam and outseam - converting pixel distances to real-world cm via a scale factor.

03 Our Solution

Built the Cloth Measurement body-sizing service (body_measurements branch). Developed a FastAPI backend with a MediaPipe-based PoseDetector that tracks key body landmarks (nose, shoulders, wrists, ankles) and a MeasurementCalculator that converts pixel landmarks into real-world garment measurements using a height-normalized scale factor, exposing stateful capture-session endpoints for front and side photo / real-time input.

Engineering approach

Engineered a stable-capture pipeline with temporal landmark smoothing, required-consecutive-frame stability checks and front/side orientation detection so measurements are only taken from confident, well-framed frames. Applied a cm-per-pixel scale factor (calibrated from real height) to compute Height, Neck, Shoulder, Chest, Waist, Hip, Arm, Inseam and Outseam sizes, and containerized the service with Docker behind an API-key gate.

04 Design Thinking

Scope reviewArchitecture mappingBuildLaunch

05 Technical Architecture

Architecture summary: Engineered a stable-capture pipeline with temporal landmark smoothing, required-consecutive-frame stability checks and front/side orientation detection so measurements are only taken from confident, well-framed frames. Applied a cm-per-pixel scale factor (calibrated from real height) to compute Height, Neck, Shoulder, Chest, Waist, Hip, Arm, Inseam and Outseam sizes, and containerized the service with Docker behind an API-key gate.

06 Technology Stack

PythonFastAPI / Uvicorn / GunicornMediaPipeOpenCVNumPy / PillowPydanticDocker

07 Development Timeline

Delivered through structured phases - discovery, design, build, integration, and launch - with iterative releases and ongoing enhancements across a ai & computer vision delivery.

08 UI Screens / Key Features Showcase

09 Before vs After

Before

Manual, disconnected workflows and limited visibility across operations.

After

An integrated, automated ai & computer vision solution with a unified experience, stronger controls and measurable efficiency.

10 Performance Metrics

99.9%System SLA
< 90msResponse Time
High EfficiencyClient Impact

11 Business Outcomes

Outcomes available on request.

12 Testimonial & Connect

Project delivered by the Dogra Technologies engineering team.

Interested in a similar solution?

Request Architecture Brief View Full Portfolio