This repository contains the code and documentation for my master's thesis, which focuses on predicting well-being from visual data using novel AI tools for face detection, tracking, re-identification, facial emotion recognition and head pose estimation.
In most cases, an object detector is not applied at every frame when performing tracking. Instead, an object detector is typically used to identify the object of interest in the first frame, and then a tracking algorithm is used to follow the object through subsequent frames.
The tracking algorithm typically uses the initial bounding box provided by the object detector and attempts to estimate the position of the object in subsequent frames based on the motion of the object and its appearance. This can involve using techniques such as optical flow, Kalman filtering, or deep learning-based approaches to track the object.
However, there may be cases where an object detector is used at regular intervals or when the tracking algorithm loses track of the object and needs to re-detect it in the scene.
In most cases, an object detector is not applied at every frame when performing tracking. Instead, an object detector is typically used to identify the object of interest in the first frame, and then a tracking algorithm is used to follow the object through subsequent frames.
The tracking algorithm typically uses the initial bounding box provided by the object detector and attempts to estimate the position of the object in subsequent frames based on the motion of the object and its appearance. This can involve using techniques such as optical flow, Kalman filtering, or deep learning-based approaches to track the object.
However, there may be cases where an object detector is used at regular intervals or when the tracking algorithm loses track of the object and needs to re-detect it in the scene.