Brief Description
YOLO-REACT represents a state-of-the-art deep learning application that employs the YOLO (You Only Look Once) algorithm for real-time species detection. The overarching objective of this initiative is to enhance conservation management strategies through the provision of rapid and accurate species identification, thereby facilitating more informed monitoring and intervention measures.
Our initial focus encompasses the Crown-of-Thorns Starfish (COTs), indigenous to Australian waters, known for its detrimental impact on coral reefs. In addition to COTs, we are developing an algorithm tailored for the detection of Lionfish, an invasive species predominantly found in Florida.
Expanded Description
YOLO-REACT is an advanced deep learning application that harnesses the capabilities of the YOLO (You Only Look Once) algorithm to achieve real-time species detection. In an era marked by significant ecological shifts that threaten biodiversity, the integration of technology into conservation efforts is paramount. YOLO-REACT is designed to augment conservation management by providing instantaneous species identification, thereby enabling more strategic monitoring and intervention.
The Rationale -- Addressing Ecological Imperatives
Crown-of-Thorns Starfish (COTs): Indigenous to the marine regions around Australia, the Crown-of-Thorns Starfish (COTs) is a predatory echinoderm that primarily consumes coral polyps. While COTs play a role in the natural marine ecosystem, episodic population explosions result in excessive coral predation, leading to substantial ecological repercussions. Such outbreaks diminish coral cover, reduce marine biodiversity, and compromise the equilibrium of the reef ecosystem.
Through YOLO-REACT, we endeavor to provide timely detection of COTs population spikes and document their distribution. Such early detection mechanisms can guide targeted conservation responses, potentially mitigating the severity of outbreaks and facilitating reef recovery.
Lionfish: The Lionfish, a native of the Indo-Pacific region, has emerged as an invasive species in the Atlantic waters, including Florida. The absence of natural predators in these new habitats, coupled with their prolific reproductive capabilities, has led to ecological imbalances. Lionfish not only compete with native species for resources but also introduce predation pressures that alter native marine species behaviors.
Our research extends to the development of a dedicated algorithm for Lionfish detection. By accurately mapping the proliferation of Lionfish, we aim to bolster initiatives focused on population control, thereby contributing to the restoration and maintenance of marine ecological balance.
To-do:
Remove weights from assetsRemove API calls from all notebooksUpload confusion matricesMake gifs from .mp4 filesFinish writing paperStack results
Authors: kluless/Angad; Rama.