helia S.
Junior Computer Vision Specialist
Open to offersNew to PlatformAbout
Helia S. is an AI Master's student at Shiraz University, specializing in YOLOv7 for object detection with practical experience in building and refining a weapon detection system from ground up. She manually annotated 900 images using bounding boxes in YOLO format, demonstrating a strong proficiency in Python, PyTorch, and LabelImg. Her current research focuses on Transformer-based anomaly detection in surveillance videos for her M.Sc. thesis, aiming to improve detection accuracy through innovative CNN and Transformer hybrids. During her work at the University of Tabriz, she achieved significant enhancement in firearm detection accuracy and successfully reduced false positives by implementing precise bounding boxes. Helia is skilled in data annotation, dataset preparation, and computer vision and seeks part-time remote opportunities in these areas.
Experience
Anomaly Detection
Shiraz Univ. · 2025 — PresentFocused on improving detection accuracy by designing a CNN-Transformer hybrid, validated through AUC comparison with the RTFM baseline. Enhanced model attention using multi-level temporal attention visualization. Worked on achieving real-time performance by adapting a Transformer framework for surveillance video applications.Weapon Detection
Univ. of Tabriz · 2021 — PresentAchieved precise firearm detection by manually annotating 1000 images and training models using YOLOv7. Reduced false positives through the implementation of accurate bounding boxes in YOLO format. Developed a complete pipeline with Python, PyTorch, LabelImg, and OpenCV that successfully facilitated inference.
Skills & Expertise
Education
- M.Sc. in AI & RoboticsShiraz University · 2024 — 2026
- B.Sc. in Computer EngineeringUniversity of Tabriz · 2020 — 2024
Interested in this professional?
Sign in as an employer to save this profile or invite helia S. to a job.
Sign in as an employer