Visual Object Tracking Using Siam and TensorFlow as Hybrid Model in AI
Utkarsh Dubey, Raju Barskar · Traitement du signal · 2025
Tracking object from various challenges is a key motivation in computer vision and machine learning.It is bit rigorous to fulfill all the challenge with higher level of precision in considering all frames.An immaculate approach is required to build a model for better visual object tracking.It is required to obtain the patterns in each frame for ideal model.Here the research uses Siamese Network and TensorFlow to train and build the model.Siamese Network may contain two or more identical sub-networks that can compare the input and make decision more precise.It is required to pipelining the anchors with positive and negative sources to process the model towards hybrid one for processing the corresponding image.TensorFlow helps to gather the patterns of the objects and recognize it to pertain the same till last frame.Hypothesis is tested with various benchmarks including OTB50, OTB100 and TempleColor128 that pertained better level of precision.