Yaya Huang

Key Skills

Software

CVAT

LabelImg

Top Subject Matter

No subject matter listed

Top Data Types

Image

Top Task Types

Bounding Box

Freelancer Overview

Following my undergraduate degree, I gained several years of experience in Python and Java development. From 2020 onwards, I undertook online data annotation work for up to five years. In 2023, I commenced my postgraduate studies in Artificial Intelligence. I have published three papers related to artificial intelligence. Alongside my studies, I continue to pursue my online data annotation work.

Labeling Experience

Industrial Workplace Safety & PPE Detection Project

LabelimgImageBounding Box

The project aimed to monitor safety compliance in factories and construction sites via AI. Primary tasks involved annotating objects from CCTV footage, including workers, hard hats, high-visibility vests, goggles, and face masks. Beyond bounding boxes, the task required attribute classification (e.g., distinguishing hard hat colors for role identification, verifying proper mask usage). Processed approximately 20,000 images, successfully overcoming challenges such as low resolution from distant surveillance cameras, poor lighting, and crowded scenes. Maintained a delivery acceptance rate of over 97%.

The project aimed to monitor safety compliance in factories and construction sites via AI. Primary tasks involved annotating objects from CCTV footage, including workers, hard hats, high-visibility vests, goggles, and face masks. Beyond bounding boxes, the task required attribute classification (e.g., distinguishing hard hat colors for role identification, verifying proper mask usage). Processed approximately 20,000 images, successfully overcoming challenges such as low resolution from distant surveillance cameras, poor lighting, and crowded scenes. Maintained a delivery acceptance rate of over 97%.

2024 - 2025

Urban Road Object Detection and Semantic Segmentation

CVATImageBounding Box

Responsible for annotating complex urban road scene images for autonomous driving training data. Key tasks included 2D Bounding Box annotation for vehicles, pedestrians, and traffic signs, as well as pixel-level Polygon segmentation for lane lines and drivable areas. Processed approximately 15,000 images during the project scope. Adhered to strict quality guidelines, handling occluded and truncated objects precisely, maintaining an accuracy rate (IoU) of over 98%. Participated in multiple QA rounds and assisted in refining the annotation guideline documentation.

Responsible for annotating complex urban road scene images for autonomous driving training data. Key tasks included 2D Bounding Box annotation for vehicles, pedestrians, and traffic signs, as well as pixel-level Polygon segmentation for lane lines and drivable areas. Processed approximately 15,000 images during the project scope. Adhered to strict quality guidelines, handling occluded and truncated objects precisely, maintaining an accuracy rate (IoU) of over 98%. Participated in multiple QA rounds and assisted in refining the annotation guideline documentation.

2023 - 2024

Education

Work History