Auto-FishID

AI Quick Summary

Stream Ocean’s solution addresses two critical challenges: the lack of standardised ocean biodiversity metrics and the absence of effective data acquisition techniques. Driven by our mission to educate our peers and empower industry and science, we are bridging the gap between data, communication, and actionable insights. By leveraging new monitoring technologies, such as advanced underwater cameras empowered by AI software, Stream Ocean is poised to revolutionise the way we interact with and protect our oceans, ensuring a sustainable future for all.   In this project, Stream Ocean will collaborate with Siyi Zhang, a Masters candidate at ETH Zurich, to work on  our AI-based fish recognition software.  After working with Siyi for four months, Stream Ocean will be able to detect “fish” in any video we process, from freshwater, marine, or lab environments.  Her software development work would help us to improve our products and generate revenues from existing and future clients. Marine projects suffer from the absence of universally accepted biodiversity metrics and efficient data acquisition methods. This data deficiency poses a grave challenge for businesses seeking to assess their oceanic impact, significantly impeding effective decision-making and efficient resource allocation.    Stream Ocean has established a pioneering product in the field of marine biodiversity monitoring, combining cutting-edge technology with business viability. By developing state-of-the-art hardware, AI-driven analysis, and standardised metrics, we are revolutionising marine biodiversity monitoring, delivering near real-time data and engaging content, enabling our clients to achieve sustainability goals while engaging clients with their activities.   To ensure future accountability, continual data improvement, and rapid, robust product deployment, Stream Ocean’s core software infrastructure is fully cloud-based. All video data storage and analysis occurs in the cloud. This means data analyses can be both independently verified, and also improved, as we produce more training data for our AI models. This increases the quality and value of both our historical and current data product offerings over time.  Our cloud infrastructure is robust and scalable, relying on a modular architecture that allows key components – such as metrics, AI, or social media outputs – to be added or replaced as methodologies and technologies improve.  In this project, Stream Ocean will collaborate with Siyi Zhang, a Masters candidate at ETH Zurich, to establish a new AI-based fish recognition software.  After working with Siyi for four months, Stream Ocean will be able to detect “fish” in any video we process, from freshwater, marine, or lab environments.  A grant from IBSDF would allow us to implement a new fish detector, thereby producing more accurate data and better servicing our existing and future clients.  Her work would be promptly integrated into a project that we recently implemented in Palau, funded by the Asian Development Bank (ADB).

Project Idea Metadata

Project Idea Name: Auto-FishID
Date: 11/11/2025 3:39:57 PM
Administrators:

Project Idea Description

Marine projects suffer from the absence of universally accepted biodiversity metrics and efficient data acquisition methods. This data deficiency poses a grave challenge for businesses seeking to assess their oceanic impact, significantly impeding effective decision-making and efficient resource allocation.

Stream Ocean has established a pioneering product in the field of marine biodiversity monitoring, combining cutting-edge technology with business viability. By developing state-of-the-art hardware, AI-driven analysis, and standardised metrics, we are revolutionising marine biodiversity monitoring, delivering near real-time data and engaging content, enabling our clients to achieve sustainability goals while engaging clients with their activities.

To ensure future accountability, continual data improvement, and rapid, robust product deployment, Stream Ocean’s core software infrastructure is fully cloud-based. All video data storage and analysis occurs in the cloud. This means data analyses can be both independently verified, and also improved, as we produce more training data for our AI models. This increases the quality and value of both our historical and current data product offerings over time. Our cloud infrastructure is robust and scalable, relying on a modular architecture that allows key components – such as metrics, AI, or social media outputs – to be added or replaced as methodologies and technologies improve.

In this project, Stream Ocean will collaborate with Siyi Zhang, a Masters candidate at ETH Zurich, to establish a new AI-based fish recognition software. After working with Siyi for four months, Stream Ocean will be able to detect “fish” in any video we process, from freshwater, marine, or lab environments. A grant from IBSDF would allow us to implement a new fish detector, thereby producing more accurate data and better servicing our existing and future clients. Her work would be promptly integrated into a project that we recently implemented in Palau, funded by the Asian Development Bank (ADB).

In this project, Stream Ocean will collaborate with Siyi Zhang, a Masters candidate at ETH Zurich, to work on our AI-based fish recognition software. After working with Siyi for four months, Stream Ocean will be able to detect “fish” in any video we process, from freshwater, marine, or lab environments. Her software development work would help us to improve our products and generate revenues from existing and future clients.