لطفا منتظر بمانید ...
0% Complete
صفحه اصلی
/
یازدهمین کنفرانس بین المللی کنترل ، ابزار دقیق و اتوماسیون
A Fast and Semi-Automatic Video Annotation Method Using Image Tracking for Deep Learning Training
نویسندگان :
Mahdi Pourkerman
1
Seyed Vahab Shojaedini
2
1- دانشگاه آزاد واحد الکترونیک
2- پژوهشکده مهندسی برق و کامپیوتر سازمان پژوهشهای علمی و صنعتی ایران
کلمات کلیدی :
Annotation،Deep Learning
چکیده :
Accurate annotation of moving objects in videos is a critical yet time-consuming task in many computer vision projects. Traditional frame-by-frame methods demand extensive effort and high user precision. This paper proposes a semi-automatic and efficient approach for annotating moving objects by leveraging image-based tracking techniques. Users annotate only keyframes, while intermediate frames are automatically inferred. Animation tools such as interpolation allow for rapid correction of tracking errors. The method produces output compatible with standard formats like YOLOv8 (aligned with YOLOv5 and above), and can be adapted to COCO and other formats, making it suitable for both online and offline applications. Implementation is done via a custom script developed in Adobe After Effects [11], enabling automated extraction of object positions and generation of annotations. Experimental evaluations were conducted on a dataset tracking cattle movement in videos, demonstrating both time efficiency and spatial consistency improvements over traditional methods.
لیست مقالات
لیست مقالات بایگانی شده
یک روش یادگیری عمیق با نظارت مبتنی بر حافظه طولانی کوتاه مدت برای مدلسازی حسگر های نرم
شیدک یزدانی - سید جلیل ساداتی رستمی - سپیده عیسیپور
Model Predictive Control of Three-Phase Four-Leg MMC-Based STATCOM Under Nonlinear Load Phase Switching Conditions
Milad Samady Shadlu
Dissipativity-based Stability Analysis of Power Networks with Uncertain Interconnections
Narges Rezaei Kookhdan - Hamid Reza Koofigar
Formation of Second-order Multi-agent Systems with Link Quality Aware Control
Ali asghar Rajabi - Iman Izadi - Majid Nabi
Application of NSGA-II in Channel Selection of Motor Imagery EEG Signals with Common Spatio-Spectral Patterns in BCI Systems
M. Moein Esfahani - Hossein Sadati
State-Dependent LMI-Based Controller and Observer Design for Constrained Boiler-Turbine-Generator Using Pseudo-Linearization Technique
Nasrin Kalamian - Atiyeh Keshavarz-Mohammadiyan
Membership Function Dependent Stability Analysis of Fuzzy Large-Scale Systems with Unknown Interconnection Terms Using Fuzzy Lyapunov Function
Mohammadreza Soltanian - Mohammad Sarbaz - Iman Zamani
Investigation in Avoiding Ferroresonance in PT Circuits using Shunt Capacitor
Hossein Tahami fard - Saeed Seyedtabaii
A Novel Cyber-Secure Algorithm in Cooperative Vehicles Platoon
Abolfazl Saadati Moghadam - Mohammad Haeri
Assessing Pediatric Sleep Apnea-Hypopnea Severity: Analyzing SpO2 Signals Spectrograms with Inception V3 Model
Erfan Mortazavi - Bahram Tarvirdizadeh - Khalil Alipour - Mohammad Ghamari
بیشتر
ثمین همایش، سامانه مدیریت کنفرانس ها و جشنواره ها - نگارش 44.9.5