Voxel51 Filtered Views Newsletter - July 19, 2024

#computervision #datascience #machinelearning #ai

This week's newsletter spotlights Voxel51's latest blogs, key resources, upcoming talks, and a list of upcoming events.

📙 Good Reads

Segment Anything in a CT Scan with NVIDIA VISTA-3D

Visual AI applications have swiftly integrated into every sector, including healthcare, where precision is paramount. In critical healthcare settings, where patient outcomes hinge on accurate diagnoses, medical professionals depend on advanced medical imaging and diagnostic tools.

Daniel Gural in his blog showcases how to leverage one of NVIDIA's Medical Foundation Models, VISTA-3D, to segment CT scans automatically using FiftyOne!

Solving the AI Blindspot: Using Data to Drive Models in Automotive

Automotive AI teams face relentless pressure to deploy new and enhanced capabilities swiftly. Whether focusing on adaptive cruise control, collision avoidance, sensor obstruction detection, in-cabin distracted driver monitoring, or other critical AI-powered systems for modern and future automobiles, projects can veer off course due to blindspots in AI development.

Check out the blog by Dave Mekelburg on how leading OEMs and Tier 1 suppliers are leveraging Voxel51 solutions to align data and models with effective workflows, ensuring successful visual AI applications.

🔎Good Finds

Hugging Face Hub

You can now filter datasets by library on the Hugging Face Hub. It's now easier than ever to find the perfect dataset for your use case 💪

Since we integrated FiftyOne datasets with the Hugging Face Hub a few months ago, 50+ FiftyOne datasets have been pushed to the hub, including datasets for image classification, object detection, video classification, image captioning, VQA, and 3D tasks. And more are being shared every day!

Check out the Hugging Face Datasets Hub

Want to load a visual dataset from the 🤗 hub, or share your visual dataset with the community? Check out this blog post from Jacob Marks which will show you just how easy it is to get started.

WayveScenes101

Check out Harpreet Sahota 🥑’s post on YouTube adding WayveScenes101 into FiftyOne- a publicly available dataset from Wayve, an autonomous driving company, meant to serve as a benchmark for novel view synthesis models for driving applications.

The Dataset is a collection of high-resolution images and camera poses designed to advance research in novel view synthesis and scene reconstruction for autonomous driving applications. This unique Dataset of multi-view driving scenes is an absolute must if you're training or evaluating novel view synthesis models on realistic automotive data.