×
×
Conservation X Labs

Browse

Choose from our library of machine learning algorithms designed for wildlife and environmental detection,or create your own model with no coding.

The Sentinel device and machine learning models are designed and trained to send you information only on what matters to you. Start with your question, choose a model focused on the animal and feature you're interested in, or design and create it yourself without writing a single line of code.

The Model Marketplace

Just like an app store, browse through pre-trained models and select the algorithm that matches your questions and needs. This is an ever-growing list filled with algorithms made in the machine-learning communities, collaborators, partners, and engineers at CXL themselves.

Custom Machine-Learning Model Maker

Can't find a model that meets your needs on the marketplace? We have a built a pipeline that allows you to create your own unique models within minutes, just by answering a few questions and clicking some images of animals.

We will soon release a video with a walkthrough of creating your own model on our website right here.

Tiny, Focused Machine-Learning

The trouble with current machine learning models is they try to identify every object you can think of, and requires huge amounts of computer power, battery power, and data transfer to run. Our models are small, TINY in fact, which is why we make our models very focused on a per-application basis. Each model is limited to a max of three different animals to search for, so that we can still send you accurate information captured in near real-time, while using very little battery and reaching field deployment times of months to years at a time.

We are currently developing the ability to plug in a solar cell, allowing for your device to stay out in the field indefinitely collecting and sending data packets of information to you.

Identification technologies

Our first prototype uses a regular RGB camera to take motion-triggered images and inference them onboard the device (on the Edge). Once we have launched this prototype, we will be developing acoustic and video versions, and experimenting with incorporating thermal and multi-spectral imagery as identification technology options for your device to use.

This is some text inside of a div block.
Text Link
This is some text inside of a div block.
Text Link
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Conservation X Labs

Design

Customize your experience by setting your notification and actionable insight options. How you choose to receive and act on your data is up to you.

Once you've chosen a model or built one yourself, you have complete control over your model in the algorithm settings page.

Choose how you get your data

Text message, Slack notification, Email, whatever is most convenient, WALLFLWR's notifcations are flexible and work with your preferences.

Choose between a daily/weekly digest or notifications as they happen in the wild! Notifications only send you infomraton that's important to you, regarding your device, models, and events that are captured.

Choose how you act on your data

Send a text, facebook message, twitter post, Instagram DM, slack message, email, ...

Now that you have your notification, what do you want to do with it? The list goes on and on and on for choices using If This Then That (IFTTT) technology. This list is ever growing, and after launching our first prototyping, we will be development easy plug-in integrations for third-party systems and technology to 'activate' when you get a notification. For example, sounding an alarm on-site by the device with third-party speakers plugged in.

Choose what data you keep

A lot of data collected is useful, but might not necessarily be what you're looking for. Nonetheless it would be a shame to miss the chance to spot a rare endangered animal you just weren't asking about. Our algorithms trigger image capture on motion, analyze whether the animal matches those on your model, and sends you the relevant information if there is a match. However, if the device spots an animal that ISN'T a match with one on your model, the image will still be stored on the device's SD card available to you when you pick up the device. We will still notify you that an image of an unknown animal was captured and stored for you.

Design

Customize your experience by setting your notification and actionable insight options. How you choose to receive and act on your data is up to you.

This is some text inside of a div block.
Text Link
This is some text inside of a div block.
Text Link
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Conservation X Labs

Launch

We'll take it from here and handle installing the model onto your device. Give us 24 hours and we'll let you know when you're up and running!

The final step is to deploy your model onto a device. This is done over-the-air (OTA) so the device doesn't have to be near you to install new models or update current ones.

OTA installation

Once your model is chosen or created, and you have set your algorithm preferences in the settings page, you're ready to go! OTA installation makes it far easier to deploy and install new models, as well as update existing ones, on your devices or public ones.

How raw data is stored

All raw data captured is stored in an onboard SD card in the device itself. Only small data packets with important bite-size information are sent to you, so you don't have to wait to retrieve your device to know what's out there. But the raw data will be available for you on that SD card once you decide to pick up your device from deployment in the wild.

Actionable insights

We have made your near real-time notifications actionable. Once you receive information from your device on deployment, you have the choice on what to do next (you can set these options in your algorithm's settings page). Want to alert someone, sound an alarm, forward information along or post to social media? All these choices and more are possible in your algorithm settings before installing your model on device (or updating models that are already installed).

Don't have a device?

Whether you don't have your own device, the right sensors or just want to try out our system before owning your own device, members of the community can make their own devices public for you to rent and install models onto for yourself. There's no trouble with device deployment, maintenance, or retrieval, as long as there is a public device in your location of interest.

If you own a device and want to make it public, you will receive a small compensation every time another user rents your device for their own model(s).

This rent-a-device feature is in development once we have launched our prototype.

Making networks of other devices smarter

If you own another environmental data collector (for example, camera trap), there's no need to replace it with our device. The WALLFLWR can fit into whatever device ecosystem you might already have.

Place a Wifi enabled SD card into your current devices (just replace the SD card that is already inside), and our device i able to collect images (and soon sound recordings) that are captured by your current network of device(s) as long as they are in range.

Once collected, the WALLFLWR device can run its models on the data scraped from your current devices to find anything of importance to send to you. Our goal is to make your current devices better, not replace them.

The team trying to make your life easier

This is some text inside of a div block.
Text Link
This is some text inside of a div block.
Text Link
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Browse

Click to learn more
Photo by Thomas Kelley

Design

Click to learn more
Photo by Sacha Styles

Launch

Click to learn more
Photo by Eberhard Grossgasteiger
Conservation X Labs

Browse

Choose from our library of machine learning algorithms designed for wildlife and environmental detection,or create your own model with no coding.

The Sentinel device and machine learning models are designed and trained to send you information only on what matters to you. Start with your question, choose a model focused on the animal and feature you're interested in, or design and create it yourself without writing a single line of code.

The Model Marketplace

Just like an app store, browse through pre-trained models and select the algorithm that matches your questions and needs. This is an ever-growing list filled with algorithms made in the machine-learning communities, collaborators, partners, and engineers at CXL themselves.

Custom Machine-Learning Model Maker

Can't find a model that meets your needs on the marketplace? We have a built a pipeline that allows you to create your own unique models within minutes, just by answering a few questions and clicking some images of animals.

We will soon release a video with a walkthrough of creating your own model on our website right here.

Tiny, Focused Machine-Learning

The trouble with current machine learning models is they try to identify every object you can think of, and requires huge amounts of computer power, battery power, and data transfer to run. Our models are small, TINY in fact, which is why we make our models very focused on a per-application basis. Each model is limited to a max of three different animals to search for, so that we can still send you accurate information captured in near real-time, while using very little battery and reaching field deployment times of months to years at a time.

We are currently developing the ability to plug in a solar cell, allowing for your device to stay out in the field indefinitely collecting and sending data packets of information to you.

Identification technologies

Our first prototype uses a regular RGB camera to take motion-triggered images and inference them onboard the device (on the Edge). Once we have launched this prototype, we will be developing acoustic and video versions, and experimenting with incorporating thermal and multi-spectral imagery as identification technology options for your device to use.

This is some text inside of a div block.
Text Link
This is some text inside of a div block.
Text Link
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Conservation X Labs

Design

Customize your experience by setting your notification and actionable insight options. How you choose to receive and act on your data is up to you.

Once you've chosen a model or built one yourself, you have complete control over your model in the algorithm settings page.

Choose how you get your data

Text message, Slack notification, Email, whatever is most convenient, WALLFLWR's notifcations are flexible and work with your preferences.

Choose between a daily/weekly digest or notifications as they happen in the wild! Notifications only send you infomraton that's important to you, regarding your device, models, and events that are captured.

Choose how you act on your data

Send a text, facebook message, twitter post, Instagram DM, slack message, email, ...

Now that you have your notification, what do you want to do with it? The list goes on and on and on for choices using If This Then That (IFTTT) technology. This list is ever growing, and after launching our first prototyping, we will be development easy plug-in integrations for third-party systems and technology to 'activate' when you get a notification. For example, sounding an alarm on-site by the device with third-party speakers plugged in.

Choose what data you keep

A lot of data collected is useful, but might not necessarily be what you're looking for. Nonetheless it would be a shame to miss the chance to spot a rare endangered animal you just weren't asking about. Our algorithms trigger image capture on motion, analyze whether the animal matches those on your model, and sends you the relevant information if there is a match. However, if the device spots an animal that ISN'T a match with one on your model, the image will still be stored on the device's SD card available to you when you pick up the device. We will still notify you that an image of an unknown animal was captured and stored for you.

Design

Customize your experience by setting your notification and actionable insight options. How you choose to receive and act on your data is up to you.

This is some text inside of a div block.
Text Link
This is some text inside of a div block.
Text Link
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Conservation X Labs

Launch

We'll take it from here and handle installing the model onto your device. Give us 24 hours and we'll let you know when you're up and running!

The final step is to deploy your model onto a device. This is done over-the-air (OTA) so the device doesn't have to be near you to install new models or update current ones.

OTA installation

Once your model is chosen or created, and you have set your algorithm preferences in the settings page, you're ready to go! OTA installation makes it far easier to deploy and install new models, as well as update existing ones, on your devices or public ones.

How raw data is stored

All raw data captured is stored in an onboard SD card in the device itself. Only small data packets with important bite-size information are sent to you, so you don't have to wait to retrieve your device to know what's out there. But the raw data will be available for you on that SD card once you decide to pick up your device from deployment in the wild.

Actionable insights

We have made your near real-time notifications actionable. Once you receive information from your device on deployment, you have the choice on what to do next (you can set these options in your algorithm's settings page). Want to alert someone, sound an alarm, forward information along or post to social media? All these choices and more are possible in your algorithm settings before installing your model on device (or updating models that are already installed).

Don't have a device?

Whether you don't have your own device, the right sensors or just want to try out our system before owning your own device, members of the community can make their own devices public for you to rent and install models onto for yourself. There's no trouble with device deployment, maintenance, or retrieval, as long as there is a public device in your location of interest.

If you own a device and want to make it public, you will receive a small compensation every time another user rents your device for their own model(s).

This rent-a-device feature is in development once we have launched our prototype.

Making networks of other devices smarter

If you own another environmental data collector (for example, camera trap), there's no need to replace it with our device. The WALLFLWR can fit into whatever device ecosystem you might already have.

Place a Wifi enabled SD card into your current devices (just replace the SD card that is already inside), and our device i able to collect images (and soon sound recordings) that are captured by your current network of device(s) as long as they are in range.

Once collected, the WALLFLWR device can run its models on the data scraped from your current devices to find anything of importance to send to you. Our goal is to make your current devices better, not replace them.

The team trying to make your life easier

This is some text inside of a div block.
Text Link
This is some text inside of a div block.
Text Link
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Browse

Photo by Thomas Kelley

Design

Photo by Sacha Styles

Launch

Photo by Eberhard Grossgasteiger