Daniel Lullulangi

Key Skills

Software

CVAT

Label Studio

Roboflow

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Top Task Types

Bounding Box

Classification

Object Detection

Segmentation

Tracking

Freelancer Overview

I have hands-on experience in data labeling, particularly for image datasets used in computer vision projects. My work has involved annotating hundreds of images using tools such as CVAT, Label Studio, and Roboflow. I understand the importance of clear labeling guidelines, label consistency, and quality assurance in building reliable AI models. I’m especially interested in contributing to datasets that support machine learning in real-world applications, and I aim to produce clean, structured training data that meets project standards.

Labeling Experience

Soccer Player Velocity and Total Distance Covered Tracking

CVATImageBounding BoxTracking

This project aims to estimate two common parameters in soccer player tracking, speed and total distance covered (TDC), using optical tracking (OT), bridging OT system in sports performance. The OT systems are built using tracking-by-detection paradigm, combining object detector (YOLOv10/11/12) and tracking algorithms (SORT, DeepSORT, and ByteTrack). The trainings of the object detection system are done using SoccerNet dataset. The tracking is evaluated on manual-assigned camera dataset. The estimation of speed and TDC is performed on Alfheim dataset, using MAE, RMSE, and r as the metrices. The manual-assigned camera dataset is built using CVAT for tracking datasets.

This project aims to estimate two common parameters in soccer player tracking, speed and total distance covered (TDC), using optical tracking (OT), bridging OT system in sports performance. The OT systems are built using tracking-by-detection paradigm, combining object detector (YOLOv10/11/12) and tracking algorithms (SORT, DeepSORT, and ByteTrack). The trainings of the object detection system are done using SoccerNet dataset. The tracking is evaluated on manual-assigned camera dataset. The estimation of speed and TDC is performed on Alfheim dataset, using MAE, RMSE, and r as the metrices. The manual-assigned camera dataset is built using CVAT for tracking datasets.

2025 - 2025

AI-based Automated QC with Machine Vision for 3D Bioprinting

Label StudioImageSegmentation

This project is a collaborative approach by CGBio, Co. Ltd. and Universitas Indonesia where the aim of the project is to develop control system with AI-vision-based for QC process in 3D bioprinting process. Manual labeling for semantic segmentation task is performed by students in AI-topic class using LabelStudio on November 2024 to provide ground truth as the machine vision is being trained. The results are treated as decision-support system to estimate whether the result of the bioprinting process is defective or not.

This project is a collaborative approach by CGBio, Co. Ltd. and Universitas Indonesia where the aim of the project is to develop control system with AI-vision-based for QC process in 3D bioprinting process. Manual labeling for semantic segmentation task is performed by students in AI-topic class using LabelStudio on November 2024 to provide ground truth as the machine vision is being trained. The results are treated as decision-support system to estimate whether the result of the bioprinting process is defective or not.

2024 - 2024

Education

Work History