About This Project
Aquatic invasive species, such as red swamp crayfish in Japan and Europe, have significant impacts on biodiversity, human health and the economy. Quantifying the distribution of invasive species is a critical step in preventing and reducing their invasion, however, effective methods have not yet been established. Here, we aim to provide a low-cost framework for monitoring aquatic invasive species combining underwater imaging and machine learning based detection.
Ask the Scientists
Join The DiscussionWhat is the context of this research?
Biological invasions are a global problem with significant impacts on biodiversity, human health, and the economy. For example, the red swamp crayfish (Procambarus clarkii) is an invasive species in Japan that threatens local biodiversity. Quantifying the distribution of invasive species is an critical step in preventing the invasion. Current efforts rely on sampling, which requires human resources, or eDNA, which requires special equipment, thus lacking a scalablity. Recent advances in low-cost microscopy, such as the PlanktoScope and ESPressoscope, have provided scalable solutions for monitoring microbial diversity in situ. Inspired by these techniques, we aim to develop frugal monitoring tools for macroscopic biodiversity, specifically invasive species in aquatic environments.
What is the significance of this project?
Our project aims to establish a low-cost monitoring framework for invasive aquatic species by combining underwater imaging and machine learning (ML). We specifically target the invasive crayfish Procambarus clarkii, especially its juvenile stage, which has a significant impact on growth and dispersal. ML algorithms will be trained on datasets of different environmental and behavioral patterns of juvenile crayfish. The model will detect the target species in real time in images captured by the device. This approach provides quantification of juvenile populations, which are often overlooked but critical for controlling the spread of invasive crayfish. In addition, the low-cost equipment and flexibility of the ML model will allow us to extend it to other aquatic invasive species.
What are the goals of the project?
The goal of our project is to develop a scalable framework for monitoring aquatic invasive species, including crayfish. We have already developed a preliminary hardware prototype, but have not yet been finalized it to detect invasive crayfish. First, we will generate a dataset of juvenile crayfish using ornamental crayfish, and build a base of ML model. Next, we will conduct field tests to optimize the device and the ML model. We will then expand field testing in Germany, India, the US and Japan where team members are located, to establish the whole pipeline. Finally, we plan to document the protocol of the hardware and ML model as open source.
Budget
To run our project, we need a budget for device development and ML model development, ornamental crayfish for dataset creation, and fieldwork expenses. For hardware development, we will use Raspberry Pi, cameras, LTE modules, and transparent boxes. For ML model development, we need access to a cloud server with GPUs. Model learning process require a dataset for juvenile crayfish, which can be obtained by recording ornamental crayfish. In addition, our team will conduct field work in several locations in Germany, India, the United States, and Japan.
Endorsed by
Project Timeline
We plan our project as follows. We will use 2 months to generate a dataset of juvenile crayfish using ornamental crayfish, and develop the basic ML model. For the next 4 months, we will conduct fieldwork testing to optimize the hardware and ML model for real-time monitoring. For the next 4 months, we will finalize the whole framework with extended fieldwork in Germany, India, the United States and Japan. For the last 2 months, we document and publish our protocol as open source.
Jul 01, 2024
Dataset generation of juvenile crayfish with ornamental crayfish and ML model development
Jul 18, 2024
Project Launched
Nov 01, 2024
Optimization of device and ML model with fieldwork
Feb 01, 2025
Final adjustment of device and ML model with fieldwork
May 01, 2025
Documentation and publication as open source
Meet the Team
Affiliates
Team Bio
We are a cross-disciplinary team with members from around the world. Our project started in the Frugal Science class at Stanford University.
Shuzo Kato
I am a PhD student at Cluster of Excellence Physics of Life, TU Dresden, Germany, studying physical rules in the biological organization from cells to ecosystems.
Fascinated by the beautiful nature of life, I have worked on various biological problems using multidisciplinary approaches from engineering, physics, and biology. At the same time, I have always wondered how my scientific skills can approach the global problems around us. The idea for this project started in the frugal science class at Stanford, where I found that frugal tools can potentially solve the problem of biological invasions. Growing up in Japan, I took invasive alien species such as red swamp crayfish for granted and saw that there was no effective solution. I believe that our frugal approach can provide practical solutions to this problem.
Gayathri P S
Innovator
A keen observer of Machine Behaviour, I am a recent BS-MS Graduate in Biology from IISER Tirupati, India. Aspiring to be a theoretical biologist in the future I am interested in morphogenesis during early vertebrate development. If not lazing around you can find me sketching, only to make weird interpretations of it later. I read Murakami and writes a little bit under the "the blue".
My interest in ecology stems from my childhood experiences of visiting multiple national parks and wildlife sanctuaries during vacations.An escape into the nature for the adults,it instilled in me I believe something I truly cherish -- a love for conservation. During my university while contributing to a project that involved analysing acoustic signature of an endemic species which was nocturnal in nature I rediscovered my passion for conservation ecology. The sheer amount of complexity and thought in addressing the question was indeed a game changer for me. I exhausted the list of electives available in ecology and evolution at university and felt the need for more and here I am.Frugal Science have restated my faith in innovating simple solutions to complex problems.
Divyanshi Srivastava
Divyanshi is an engineer majored in biotechnology, and her broader future aim is to make this planet a better place to live. Currently she is pursuing PhD in Organisational Behaviour. Her interest in environmental sciences was born through an open elective class of Prof Ravi Prakash in 2020.
Having deeper discussions on topics like energy management and solving various case studies regarding health and environmental challenges made her more empathetic towards the issues we face in the 21st Century.
Joining the Global Frugal Science cohort in 2020 under Dr. Manu Prakash of Stanford has been a turning point in her life! She has been an active member of this cohort ever since then. Getting out of the technophilic mindset and trying to solve problems we deeply care about is the major aim of the whole course.
She and her team want to carry forward the project they have been working on, the initial thrust of which has been provided through the global cohort! She looks forward to all the like-minded people who will help them realise their ideas!
Tanvi Dutta Gupta
I'm a graduate student at Stanford University, where I study community ecology, environmental justice, and science communication. I grew up across Singapore and India and have a strong interest in exploring solutions that empower place-based environmental solutions for the Global South. My research background focuses on understanding the ecological networks connecting people, animals, and their environment, from the deep sea to urban centers. I'm fascinated by understanding how tools inform our scientific results, and love exploring what's offered by different ways of knowing, including traditional ecological knowledge, eDNA, population modeling, and theoretical ecology.
Lab Notes
Nothing posted yet.
Project Backers
- 6Backers
- 114%Funded
- $4,590Total Donations
- $765.00Average Donation