HaCKRS
The first fish-school monitoring dataset that pairs RGB imagery with sonar, containing more than 105,000 time-aligned frame pairs. It provides sonar geometry, coordinate information and planar registration, and supports the development of methods for fish detection, tracking and counting.
Timeline (last 90 days)
The researchers released HaCKRS, the first paired RGB and sonar dataset for fish-school monitoring.
The HaCKRS dataset contains more than 105,000 time-aligned frame pairs, totaling about 3.3 hours, covering four distance and orientation configurations.
The HaCKRS dataset does not provide task-specific fish body annotations; it focuses on paired observations and calibration information.
The HaCKRS dataset pairs imaging-sonar recordings from the Garmin LiveScope with top-down RGB video of koi.
The HaCKRS dataset provides per-frame metadata, sonar geometry and coordinate information, synchronized data for the four configurations, and planar registration for the three horizontal configurations.
The paper explicitly states that the goal of the HaCKRS dataset is to support the development of fish-school detection, tracking, and counting methods and to explore the potential of consumer-grade sonar for ecological monitoring.
HaCKRS's paired structure provides a natural experimental platform for multimodal learning, domain adaptation, and weakly supervised learning, because the same scene has both high-quality and low-quality observations.
The HaCKRS dataset has been made public, including synchronized data and registration information, on which researchers can carry out tasks such as fish-school detection, tracking, and counting.
All Facts (8)
HaCKRS's paired structure provides a natural experimental platform for multimodal learning, domain adaptation, and weakly supervised learning, because the same scene has both high-quality and low-quality observations.
50%UnverifiedThe HaCKRS dataset contains more than 105,000 time-aligned frame pairs, totaling about 3.3 hours, covering four distance and orientation configurations.
50%UnverifiedThe HaCKRS dataset does not provide task-specific fish body annotations; it focuses on paired observations and calibration information.
50%UnverifiedThe HaCKRS dataset pairs imaging-sonar recordings from the Garmin LiveScope with top-down RGB video of koi.
50%UnverifiedThe HaCKRS dataset provides per-frame metadata, sonar geometry and coordinate information, synchronized data for the four configurations, and planar registration for the three horizontal configurations.
50%UnverifiedThe paper explicitly states that the goal of the HaCKRS dataset is to support the development of fish-school detection, tracking, and counting methods and to explore the potential of consumer-grade sonar for ecological monitoring.
50%UnverifiedThe HaCKRS dataset has been made public, including synchronized data and registration information, on which researchers can carry out tasks such as fish-school detection, tracking, and counting.
50%UnverifiedThe researchers released HaCKRS, the first paired RGB and sonar dataset for fish-school monitoring.
50%