{"id":17906,"date":"2025-11-05T11:21:35","date_gmt":"2025-11-05T09:21:35","guid":{"rendered":"https:\/\/spatialworld.fi\/?page_id=17906"},"modified":"2025-11-27T10:58:19","modified_gmt":"2025-11-27T08:58:19","slug":"fme-blog-dinov3","status":"publish","type":"page","link":"https:\/\/spatialworld.fi\/en\/fme-blog-dinov3\/","title":{"rendered":"fme-dinov3"},"content":{"rendered":"[vc_row type=”in_container” full_screen_row_position=”middle” column_margin=”default” column_direction=”default” column_direction_tablet=”default” column_direction_phone=”default” scene_position=”center” text_color=”dark” text_align=”left” row_border_radius=”none” row_border_radius_applies=”bg” overflow=”visible” overlay_strength=”0.3″ gradient_direction=”left_to_right” shape_divider_position=”bottom” bg_image_animation=”none”][vc_column column_padding=”no-extra-padding” column_padding_tablet=”inherit” column_padding_phone=”inherit” column_padding_position=”all” column_element_direction_desktop=”default” column_element_spacing=”default” desktop_text_alignment=”default” tablet_text_alignment=”default” phone_text_alignment=”default” background_color_opacity=”1″ background_hover_color_opacity=”1″ column_backdrop_filter=”none” column_shadow=”none” column_border_radius=”none” column_link_target=”_self” column_position=”default” gradient_direction=”left_to_right” overlay_strength=”0.3″ width=”1\/1″ tablet_width_inherit=”default” animation_type=”default” bg_image_animation=”none” border_type=”simple” column_border_width=”none” column_border_style=”solid”][vc_column_text css=”” text_direction=”default”]\n

<\/h1>\n

🦎Breathing New Life into Raster Data with AI Embeddings and FME🦖<\/h1>\n[\/vc_column_text][vc_column_text css=”” text_direction=”default”]Geospatial imagery and other raster datasets often sit underutilized in archives, rich with potential insights that are hard to extract. Today, a class of AI models called embedding models<\/strong> is changing that. By using embeddings, numerical representations of data that capture its essential patterns, we can unlock information in current and legacy raster data that was previously inaccessible.<\/p>\n

In the past, embeddings were mostly an internal step in classification models. For example, a neural network might learn features that help distinguish a cat from a dog, but those features were not typically used directly. In the geospatial domain, you usually still had to train a model on your own labelled data to get usable results.<\/p>\n

With generic models like DINOv3<\/strong>, you can now compute embeddings directly on your own imagery, on your own machines. You can still fine-tune them if needed, but as we\u2019ll see, you can already do a lot just by using the pretrained model \u201cas is\u201d. Of course, you can also build on top of them, this is why Meta call it a vision foundation model.<\/p>\n[\/vc_column_text][image_with_animation image_url=”17955″ image_size=”full” max_width=”100%” max_width_mobile=”default” animation_type=”entrance” animation=”None” animation_movement_type=”transform_y” hover_animation=”none” alignment=”center” border_radius=”none” box_shadow=”none” image_loading=”default”][vc_column_text css=”” text_direction=”default”]\n

Image from https:\/\/ai.meta.com\/dinov3\/<\/em><\/p>\n

With the flexibility of Safe Software\u2019s FME<\/strong> (Feature Manipulation Engine) as a low-code integration platform, non-coders can harness these advanced AI capabilities. In this article, we\u2019ll explain embeddings in simple terms and walk through three real use cases where combining an embedding model with FME gives raster datasets a new life. Each scenario is actionable, with data and processes available for readers to experiment with themselves (see the download link at the end).<\/p>\n

<\/h2>\n

What Are Embeddings?<\/h2>\n

In the context of machine learning, embeddings<\/strong> are a way to convert complex data (like an image or text) into a list of numbers (a feature vector<\/em>) that captures meaningful characteristics of that data. You can think of an embedding as a compact summary or fingerprint of the original content.<\/p>\n

For images, an embedding might be a vector of a few hundred or a thousand numbers that collectively represent things like shapes, textures and structures in the patch, but not in an explicit human-readable way. The key idea is that similar data will have similar embeddings<\/strong>. If two image patches both contain dense forest, an embedding model will generate vectors for them that are numerically close to each other. The model doesn\u2019t just look at individual pixels, but also at patterns and context.<\/p>\n

This technique compresses vast amounts of visual information into a smaller set of features that represent the data\u2019s meaningful semantics. The power of embeddings is that they make it easy for computers to compare and search content. Instead of comparing raw pixel values (which is impractical), we compare the embeddings. This unlocks operations such as:<\/p>\n