 The speaker to start us off is Tom Y presenting this EL. AI, the citizen science based approach to aggregating images of fish and providing an online solution with open source and open access to users globally. Welcome to the official that AI presentation on how we are enabling official recognition. Meet our team Tom and Jane, why founders of the white foundation and main supporters of the official that AI project. They're also the creators of the mobile sport fishing platform fish angler. The biologist Christina activator, Bradwell, Kimberly, Dennis Shaw, Santiago, Tricati and guy a row. A marketing department stereo staff ski and joy flexor. Our current partnerships are with the wife foundation and fish angler.com. This project was originally slated to be part of fish angler. However, Tom and Jane why found it would be more beneficial to the science community if this was offered as an open source tool. The guy Harvey ocean foundation, Microsoft code ahead and conservation. Our mission, our goal is to accelerate the research and development of a highly accurate fish identification technology using AI to recognize all fish species worldwide. Our vision is to act as a central hub for the combination of various fish image data set into a single open source AI and already labeled fish image data set that will contribute to the creation of official recognition. Our goals to provide a specialized web portal for the collection and ML tagging a fish images to publish the world's largest labeled fish image database for AI ML for commercial and non commercial use. And to seek the development of a highly accurate AI model that can identify species worldwide. We are accomplishing our mission to two main actions. Our open source labeled fish image database that will be a tool for others to build AI technology as well. And the citizen science portal that will be a web based portal that allows users to upload tag images of fish species name and tag polygon. Our open source labeled fish image database came about because we noticed a limited number of labeled fish images available as open source when we began our journey. So we set out to create the world's largest. We knew that it would require a few things. It would require images from all over the world. It will require the development of a specialized website for photo repository. And it will require the ability to allow copyright data to be set for images submitted to the database, including but not limited to internal use only and not for open source data or open source friendly. Our citizen science portal would be is going to be the way we build our image database. It allows for individuals or teams to register an account. You can upload fish images easily through drag and drop, a computer upload or a CSV file upload. It has a specialized ML editor that allows for polygons to be set around each fish in the image, as well as each trait for each fish in the image and labeling polygons with fish species. The citizen science portal also allows for specialized data collection for each image, including GPS location fish species within the image and metadata. It has the ability to complete copyright data customization for each image or you can set it for a group of images. The team setting allows for the team to review images before submission to the model. And our QA system allows for the biologist to verify image prior to final submission to the model. We have a few hurdles to success. Image acquisition for less known fish and non game fish. Copyright and open source access images scraping images from social media could be very beneficial. However, it's usually against their copyright way around this would be very useful. Images that can be freely used for commercial and non commercial use. We're limited to manpower to label each image for fish identification and manpower to QA images before submission into the model. We can only do as many as the number of biologists we have. And our last hurdle to success is gaining momentum within the citizen science portal. How you can help us, we help gaining traction within fisheries, research teams and NGOs in terms of getting them involved in sharing their image databases with us in the citizen science portal. You can assist us in labeling the unlabeled fish images and or QA images that have been submitted to the model. And funding to help contribute to this effort would go to coding and web development. Building our QA team and building the fish image labeling team. Thank you for watching this watching this presentation on official AI and how we are enabling official recognition. If you would like to learn more about our projects we encourage you to visit this show that AI. If you like to email us with questions, please feel free to reach us at support at official AI. Thank you official AI for sharing and Tom, Tom, if you're there, we've got a question for you for a few questions. We normally just hand the questions around the group here and we also reach into the Q&A box if we can find one there that that's, you know, the right type of question. So please feel, feel free to hand them in there. In the search or research team use your online labeling tools for their project. Is there any cost and how easy is it for them after they've done the work to retrieve their images from your system. Yes. So currently, if you go to portal dot official dot AI sign up for an account. Basically, you can upload and also once you've tagged your images, we are using AI to automatically put bounding boxes around fish images that are uploaded. The, the actual use of the account is free, and you're able to download your images so we do wish for everybody to submit them to the overall common data set. So they can upload, edit, and then submit them to the official collection that we're calling it, and then they can also download their own data set. I've got a quick follow up before I hand over to Matt. I noticed in your presentation you said you're interested in building a data set or, you know, image sets for all fish. Do you at least try to channel people to give sets under different groupings or is it the case that you're really collecting everything that you can get. Was that to matter to Tom. Sorry, Tom, this is to you I just wanted if you've got any special angles such as commercial fish or recreational fishing or mental fish or how does it work to encourage people to maybe batch certain groups of fish. So we've created a rich set of meta data that can be attached to the images as they're uploaded. So we have specialized tags to specify if it's like an x-ray of a fish, if it's fossil, if it's underwater or above water. We have segmentation between the different types of data that you upload. Okay, Matt, have you got a question for Tom? Yes, so can you batch upload large sets of photos? Yes, you can. We have a CSV upload capability that allows you to upload about 10,000 images per batch. If I could just also introduce you to Anton Ellenbrook for our knowledge information and knowledge management section in FAO's Fisheries Division. Anton, have you got any question that might be relevant to Tom's initiative, which is trying to use citizens to upload imagery. I'm sure you've seen this time elsewhere. Yes, thanks for the presentation first. And so my question would be now in the terms of the meta data management and meta data schemas. Do you see international development taking roots where people are able to share through the meta data access to their data and also manage control and segment the data collection, image collection. So we're not using a standard per se on the export, you can export your data in a cocoa file, but we're not built in with an interface to other platforms at this time. Okay. That's already a great collection. Thanks very much. Glad we could get you on. You look like you're sitting on a nice boat there. Well, congratulations to you. Thank you very much for the invite to the conference today. You're very welcome.