{"podcast":{"title":"Satellite image deep learning","slug":"satellite-image-deep-learning-5989248","podcast_index_feed_id":5989248,"rss_url":"https://api.substack.com/feed/podcast/1186793.rss","website_url":"https://www.satellite-image-deep-learning.com/podcast","image_url":"https://substackcdn.com/feed/podcast/1186793/160a520ab8f4689dddcc760f14d20aa3.jpg","author":"Robin Cole","episode_count":46,"summary":"Newsletter on deep learning with satellite & aerial imagery","last_synced_at":"2026-06-24T18:18:37.974658+00:00","page_url":"https://stenobird.com/podcast/satellite-image-deep-learning-5989248"},"episode":{"title":"State Of The Art Object Detection","slug":"state-of-the-art-object-detection","published_at":"2026-02-04T10:26:06+00:00","page_url":"https://stenobird.com/podcast/satellite-image-deep-learning-5989248/state-of-the-art-object-detection","show_page_url":"https://stenobird.com/podcast/satellite-image-deep-learning-5989248","url":"https://www.satellite-image-deep-learning.com/p/state-of-the-art-object-detection","audio_url":"https://api.substack.com/feed/podcast/186176069/8df587cb135c0e80e027929c4291a664.mp3","summary":"In this episode I sat down with Isaac to discuss RF-DETR, a new state-of-the-art family of real-time object detection and segmentation models from Roboflow. We cover the motivation for building models that are not just accurate but also fast, cost-efficient, and deployable across diverse hardware and data regimes, and why moving beyond fixed architectures is key to achieving that. Isaac explains how RF-DETR combines strong foundation backbones like DINOv2 with efficient neural architecture search to unlock novel speed–accuracy trade-offs, including dropping decoder layers and queries after training. We also discuss the model’s strong transfer performance on domains far from COCO, the introduction of a memory-efficient instance segmentation head, and the team’s unusually rigorous benchmarking approach, before closing on the challenges of open-source research and upcoming improvements to inference and platform integration. * 👤 Isaac on LinkedIn * 🖥️ RF-DETR on Github * 📖 Paper * 📺 Video of this conversation on YouTube Bio: Isaac Robinson is a Machine Learning Research Engineer at Roboflow. He’s worked across the field of computer vision, from real-time stereo depth estimation on household robots to biomedical research at the NIH to founding a zero shot computer vision infrastructure startup. Isaac focusses on the intersection of low latency and high performance, with the goal of helping people unlock new capabilities through vision. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.satellite-image-deep-learning.com","meta_description":"In this episode I sat down with Isaac to discuss RF-DETR, a new state-of-the-art family of real-time object detection and segmentation models from Roboflo…","key_points":[],"chapters":[],"topics":[],"duration_seconds":1818,"processing_state":"not_requested","actions":[{"name":"request_transcript","method":"POST","url":"https://stenobird.com/v1/public/podcasts/satellite-image-deep-learning-5989248/episodes/state-of-the-art-object-detection/transcription-requests","description":"Idempotently request low-priority transcript generation for this episode."},{"name":"read_markdown","method":"GET","url":"https://stenobird.com/podcast/satellite-image-deep-learning-5989248/state-of-the-art-object-detection.md","description":"Read the agent-friendly Markdown representation of this episode resource."}]}}