Related Experiment Video
Updated: Jun 28, 2026
10:56
Long-term Behavioral Tracking of Freely Swimming Weakly Electric Fish
Published on: March 6, 2014
Tilapia Feeding Behavior Image Dataset: A Benchmark Resource for Automated Feeding Intensity Recognition in
Shahbaz Gul Hassan1,2, Yin Hang3, Murtaza Hasan4
1College of Bigdata and Internet, Shenzhen Technology University, Shenzhen, 518118, China.
Scientific Data
|June 26, 2026
Summary
Researchers created the Tilapia Feeding Behavior Image Dataset (TFBID-mini) to analyze Nile tilapia feeding behavior using computer vision. This dataset aids in developing automated aquaculture feeding systems.
Area of Science:
- Aquaculture
- Computer Vision
- Animal Behavior
Background:
- Real-time monitoring of fish behavior is crucial for automated aquaculture systems.
- Video-based platforms enable continuous data collection for feeding management and behavioral analysis.
Purpose of the Study:
- To introduce the Tilapia Feeding Behavior Image Dataset (TFBID-mini).
- To provide a standardized resource for evaluating AI models in aquaculture feeding behavior recognition.
Main Methods:
- Generated a dataset of 4,000 expert-annotated images from Oreochromis niloticus (Nile tilapia) feeding events.
- Captured high-definition underwater video in controlled recirculating aquaculture tanks.
- Classified images into four feeding intensity levels: None, Weak, Medium, and Strong.
Main Results:
- The TFBID-mini dataset includes images with varying illumination and quality-checked for consistency.
- The dataset represents distinct behavioral states during feeding sessions.
- It serves as a benchmark for computer vision models.
Conclusions:
- The TFBID-mini dataset facilitates the development of automated fish feeding systems.
- It supports advancements in AI-driven aquaculture monitoring and management.
- Enables objective evaluation of computer vision algorithms for fish behavior analysis.
