# Instance Flow Based Online Multiple Object Tracking

## TL;DR

This paper introduces an online multiple object tracking method that leverages instance-aware segmentation and optical flow to improve tracking accuracy in monocular videos, especially for fast-moving objects.

## Contribution

The novel approach combines instance segmentation with optical flow for object tracking, differing from traditional detection-based methods by directly modeling object shapes.

## Key findings

- Achieves a MOTA score of 32.1 on MOT 2D 2015 test set.
- Effectively tracks objects with high relative motions.
- Outperforms some previous methods in accuracy.

## Abstract

We present a method to perform online Multiple Object Tracking (MOT) of known object categories in monocular video data. Current Tracking-by-Detection MOT approaches build on top of 2D bounding box detections. In contrast, we exploit state-of-the-art instance aware semantic segmentation techniques to compute 2D shape representations of target objects in each frame. We predict position and shape of segmented instances in subsequent frames by exploiting optical flow cues. We define an affinity matrix between instances of subsequent frames which reflects locality and visual similarity. The instance association is solved by applying the Hungarian method. We evaluate different configurations of our algorithm using the MOT 2D 2015 train dataset. The evaluation shows that our tracking approach is able to track objects with high relative motions. In addition, we provide results of our approach on the MOT 2D 2015 test set for comparison with previous works. We achieve a MOTA score of 32.1.