University of Hawaiʻi at Hilo employs AI in reef fish tracking, garnering national acclaim
UH Hilo’s innovative FISHTRAC software employs AI for precise reef fish tracking.
Research paper published in Pattern Recognition, winning the editor’s choice award.
AI reduces human effort and time in image review, offering a substantial leap in efficiency.
FISHTRAC serves as an ethical alternative to invasive catch-and-release tagging methods.
The collaboration between experts from diverse disciplines drives AI’s impact on marine research.
Main AI News:
University of Hawaiʻi at Hilo has gained national recognition for its pioneering use of artificial intelligence (AI) in revolutionizing reef fish tracking. Spearheaded by computer scientist Travis Mandel and his interdisciplinary team, their FISHTRAC software has emerged as a transformative solution in the realm of marine research. This collaborative creation, involving students, alumni, and faculty, utilizes AI to meticulously trace individual fish within video footage and photographs. The result? This is a remarkable departure from the time-consuming task of manual image review. The team’s remarkable research paper, published in Pattern Recognition this past March, has not only garnered attention but also secured the prestigious editor’s choice award.
“It certainly outperforms the myriad of other algorithms we assessed in our paper, which was an extensive array,” notes Mandel, who serves as an associate professor of computer science and directs UH Hilo’s interdisciplinary data science program. While not without its imperfections, Mandel explains that AI significantly reduces the human effort and time required for reviewing extensive video content frame by frame. The traditional method of manually drawing boxes around each frame is, as Mandel puts it, “a task that seems to stretch into eternity.”
FISHTRAC software, built upon AI-based video identification, emerges as an ethical alternative to the invasive catch-and-release tagging research methods often employed. The central question guiding the research team was clear: How can AI and machine learning systems collaborate with humans to address real-world challenges? Initially called upon by environmental scientists at UH Hilo to address computer vision challenges—specifically, the consistent recognition of objects in photographs or videos—Mandel embarked on a complex journey. Teaching an AI engine to learn is a feat at the forefront of both computer science and environmental science today.
While Mandel’s expertise lay outside the realm of computer vision, the urgency for research in this field swiftly materialized. “Faculty members and graduate students from various disciplines started reaching out, saying, ‘Can you assist us with our computer vision challenges?‘” recalls Mandel.
The collaborative paper, co-authored by Mandel, UH Hilo alumni Mark Jimenez (computer science), Emily Risley (computer science), Taishi Nammoto (physics), and Rebekka Williams (mathematics), also features the contributions of UH Hilo students Meynard Ballesteros (computer science) and Bobbie Suarez (tropical conservation biology and environmental science), as well as Max Panoff, a doctoral student in electrical and computer engineering at the University of Florida. This convergence of expertise and dedication underscores the profound impact of AI on marine research, setting a new standard for innovation in the field.
The University of Hawaiʻi at Hilo’s AI-powered FISHTRAC software represents a game-changing innovation in marine research. Streamlining fish tracking and reducing the need for manual image review not only improves efficiency but also offers a more ethical approach to data collection. This breakthrough signifies the growing role of AI in addressing complex real-world challenges in various industries, including marine research, and highlights the potential for further advancements in AI-driven solutions.