HOTA: A Higher Order Metric for Evaluating Multi-object Tracking

2020/10/08 by Jonathon Luiten, Aljos̆a Os̆ep, Patrick Dendorfer +4 · 65 citations
Computer Science · #Video Surveillance and Tracking Methods #Gaze Tracking and Assistive Technology #Target Tracking and Data Fusion in Sensor Networks

paper · pdf · doi:10.1007/s11263-020-01375-2

Abstract

Multi-object tracking (MOT) has been notoriously difficult to evaluate. Previous metrics overemphasize the importance of either detection or association. To address this, we present a novel MOT evaluation metric, higher order tracking accuracy (HOTA), which explicitly balances the effect of performing accurate detection, association and localization into a single unified metric for comparing trackers. HOTA decomposes into a family of sub-metrics which are able to evaluate each of five basic error types separately, which enables clear analysis of tracking performance. We evaluate the effectiveness of HOTA on the MOTChallenge benchmark, and show that it is able to capture important aspects of MOT performance not previously taken into account by established metrics. Furthermore, we show HOTA scores better align with human visual evaluation of tracking performance.

Cited by

• AerialMind: Towards Referring Multi-Object Tracking in UAV Scenarios
• StableTrack: Stabilizing Multi-Object Tracking on Low-Frequency Detections
• SelfMOTR: Revisiting MOTR with Self-Generating Detection Priors
• Occlusion-Aware Multi-Object Tracking via Expected Probability of Detection
• LAA3D: A Benchmark of Detecting and Tracking Low-Altitude Aircraft in 3D Space
• A Tri-Modal Dataset and a Baseline System for Tracking Unmanned Aerial Vehicles
• OmniPT: Unleashing the Potential of Large Vision Language Models for Pedestrian Tracking and Understanding
• Vision-Motion-Reference Alignment for Referring Multi-Object Tracking via Multi-Modal Large Language Models
• Enhancing Multi-Camera Gymnast Tracking Through Domain Knowledge Integration
• StreetView-Waste: A Multi-Task Dataset for Urban Waste Management
• The SA-FARI Dataset: Segment Anything in Footage of Animals for Recognition and Identification
• 3D Ground Truth Reconstruction from Multi-Camera Annotations Using UKF
• PlugTrack: Multi-Perceptive Motion Analysis for Adaptive Fusion in Multi-Object Tracking
• DMSORT: An efficient parallel maritime multi-object tracking architecture for unmanned vessel platforms
• Multi-Object Tracking Retrieval with LLaVA-Video: A Training-Free Solution to MOT25-StAG Challenge
• Comparative analysis of multiple YOLO-based target detectors and trackers for ADAS in edge devices
• Zero-Shot Multi-Animal Tracking in the Wild
• Pinpointing Trigger Moment for Grounded Video QA: Enhancing Spatio-temporal Grounding in Multimodal Large Language Models
• OmniTrack++: Omnidirectional Multi-Object Tracking by Learning Large-FoV Trajectory Feedback
• Dense cotton boll counting with transformer-based video tracking and a customized phenotyping robot for data collection
• A robust visual tracking method for wind-disturbed lotus pods with reduced ID switches
• When Fish Look Alike: Tracking Identities with Dual-branch Elasticity
• WaspMOT: A Benchmark for Long-Term Multi-Object Tracking of Trichogramma Wasps
• Uniformity and deformation: A benchmark for multi-fish real-time tracking in the farming
• A Hybrid Approach for Visual Multi-Object Tracking
• GenTrack: A New Generation of Multi-Object Tracking
• Towards Physics-informed Spatial Intelligence with Human Priors: An Autonomous Driving Pilot Study
• FutrTrack: A Camera-LiDAR Fusion Transformer for 3D Multiple Object Tracking
• Multi-Camera Worker Tracking in Logistics Warehouse Considering Wide-Angle Distortion
• SAM 2++: Tracking Anything at Any Granularity
• CaptionFormer: Unified Segmentation, Tracking, and Captioning for Spatio-Temporal Objects
• EPIPTrack: Rethinking Prompt Modeling with Explicit and Implicit Prompts for Multi-Object Tracking
• MMOT: The First Challenging Benchmark for Drone-based Multispectral Multi-Object Tracking
• SVAG-Bench: A Large-Scale Benchmark for Multi-Instance Spatio-temporal Video Action Grounding
• Image-to-Video Transfer Learning based on Image-Language Foundation Models: A Comprehensive Survey
• Fast Self-Supervised depth and mask aware Association for Multi-Object Tracking
• Learning What To Hear: Boosting Sound-Source Association For Robust Audiovisual Instance Segmentation
• Motion-Aware Transformer for Multi-Object Tracking
• Learning Pseudo 3D Representation for Ego-centric 2D Multiple Object Tracking
• SynapFlow: A Modular Framework Towards Large-Scale Analysis of Dendritic Spines
• DepTR-MOT: Unveiling the Potential of Depth-Informed Trajectory Refinement for Multi-Object Tracking
• An Analysis of Kalman Filter based Object Tracking Methods for Fast-Moving Tiny Objects
• Occlusion-aware worker tracking via adaptive fusion of appearance and motion features
• Seg2Track-SAM2: SAM2-based Multi-object Tracking and Segmentation
• Motion Estimation for Multi-Object Tracking using KalmanNet with Semantic-Independent Encoding
• Online 3D Multi-Camera Perception through Robust 2D Tracking and Depth-based Late Aggregation
• Deep learning in multiple animal tracking: A survey
• NOOUGAT: Towards Unified Online and Offline Multi-Object Tracking
• See No Evil: Adversarial Attacks Against Linguistic-Visual Association in Referring Multi-Object Tracking Systems
• MVTrajecter: Multi-View Pedestrian Tracking with Trajectory Motion Cost and Trajectory Appearance Cost
• COMETH: Convex Optimization for Multiview Estimation and Tracking of Humans
• SoccerNet 2025 Challenges Results
• ByteTrack: Multi-object Tracking by Associating Every Detection Box
• SMTrack: End-to-End Trained Spiking Neural Networks for Multi-Object Tracking in RGB Videos
• FastTracker: Real-Time and Accurate Visual Tracking
• Model-based Multi-object Visual Tracking: Identification and Standard Model Limitations
• The 9th AI City Challenge
• Towards Low-Latency Tracking of Multiple Speakers With Short-Context Speaker Embeddings
• SIS-Challenge: Event-based Spatio-temporal Instance Segmentation Challenge at the CVPR 2025 Event-based Vision Workshop
• SocialTrack: Multi-Object Tracking in Complex Urban Traffic Scenes Inspired by Social Behavior
• MeMoSORT: Memory-Assisted Filtering and Motion-Adaptive Association Metric for Multi-Person Tracking
• GRASPTrack: Geometry-Reasoned Association via Segmentation and Projection for Multi-Object Tracking
• TrackOR: Towards Personalized Intelligent Operating Rooms Through Robust Tracking
• QuaDreamer: Controllable Panoramic Video Generation for Quadruped Robots
• DRWKV: Focusing on Object Edges for Low-Light Image Enhancement

Related

• MOTCOM: The Multi-Object Tracking Dataset Complexity Metric
• Tracking Every Thing in the Wild
• Local Metrics for Multi-Object Tracking
• Revisiting the details when evaluating a visual tracker
• Tracking the Trackers: An Analysis of the State of the Art in Multiple Object Tracking