IDEA-Computer Vision and Robotics

We are thrilled to announce that the IDEA Center for Computer Vision and Robotics has officially launched the DINO-X model, now available via the DDS Cloud API! As a revolutionary open-world vision model, DINO-X sets a new benchmark for universal object recognition, offering unparalleled full-scene perception capabilities: ● Optimized Long-tail Object Detection: Supports text prompts, visual prompts, and custom prompts ● Enhanced Fine-grained Object Understanding: Delivers diverse output formats, including bounding boxes, segmentation masks, key points, and descriptive texts ● Powered by a Large-scale Dataset: Built on over 100 million high-quality samples for superior model performance ● Integrated Multi-task Perception: Supports segmentation, pose estimation, region descriptions, and region-based Q&A ● Exceptional Generalization Capability: Excels in unknown scenarios and adapts seamlessly to complex environments Click here to access the documentation and start using DINO-X API: https://lnkd.in/dT2E6J83 Try Playground for DINO-X : https://lnkd.in/dKzqHh5q #AI #Deeplearning #computervision #ObjectDetection #MultimodalLargeModels #IDEAResearch

In July 2024, Grounding DINO 1.6 Pro & Edge released with larger, higher-quality datasets. Pro excels in standard detection, outperforming 1.5 Pro in specific cases. Edge optimizes for faster inference, surpassing YOLOv8 in specific scenarios. #YOLOv8 #YOLO #history #deeplearn #GenerativeAI #deeplearning

Object detection is vital in computer vision but limited by closed-set models. In Same April 2024, the open-set model T-Rex2 emerged, detecting arbitrary objects via text-visual prompts without training, excelling in dense small-object scenarios. #deeplearn #YOLO #GenerativeAI

In May 2024, the 1.5 Edge & Pro versions launched, powered by Grounding DINO. The Edge version offers fast, reliable detection with low power usage, while the Pro version leads in open-world detection performance. #machine_learning #MachineLearningAndAI #AI #CV #computervision #product #GroundingDINO #History #OSOD

Grounding DINO is a powerful open-set object detection model that detects objects based on text input. Its detection boxes can be used with models like Segment Anything and Stable Diffusion, a series of text-driven pipeline can be created. #AI #CV #computervision #product #GroundingDINO #workflow #pipeline #History #OSOD

After 2 months of continuous R&D and debugging, CountAnything was finally launched on Google Play today and successfully supports the Android version! Welcome everyone to download and experience, This AI can count! Let's set up a quick connection to CountAnything via the URL below :: https://lnkd.in/dWbcWb3m

In today's era of rapid technological development, artificial intelligence is gradually penetrating into various fields, and counting apps are no exception. This article will introduce you to the 4 best counting apps driven by artificial intelligence in detail, and show you how they use advanced technology to help you easily count any items, whether it is inventory management in business operations or item counting in daily life, which can greatly improve efficiency. 🤗 #artificialintelligencecountingapp #bestcountingtool #AItechnologyapp #operationalefficiencyimprovement #countingapp #inventorymanagementartifact #smartcountingexperiencesharing #efficientcountingmethod

T-rex Label & CountAnything participated in the product selection of ProductHunt, welcome everyone's fierce voting 🙋♂️ 🙋♀️ T-rex Label :: https://lnkd.in/d-XQQgFN CountAnything :: https://lnkd.in/dK-6mxaM #ai #annotation #developertool #ProductivityTips #producticitytools #counting #counter #countingapp #AppStore #AppleMyloveTheSeries

Grounding DINO 1.6 meets Black Myth Wukong! Detecting the Kang-Jin Loong and Black Bear Guai is a breeze! The detection of the Black Bear Guai animation also poses no challenge, and not even its shadow is overlooked! Playground experience：https://lnkd.in/dBknYdir #BlackMythWuKong #AI #GroundingDINO #computervision #objectdetection

T-Rex Label is an intelligent annotation tool built on the zero-shot detection model T-Rex2. It breaks the traditional dependency on training data, directly using visual prompts as input, significantly improving annotation efficiency! The video highlights the wide applications of T-Rex Label in traffic annotation, including parking space detection, vehicle identification, traffic flow detection, road waterlogging detection, road pothole detection, road obstacle detection, pedestrian flow counting, airport management and vessel traffic counting. With just a visual prompt, it enables one-click annotation, AI-assisted corrections, and automatic counting, significantly improving annotation efficiency and providing high-quality datasets for traffic model training! Try out this powerful AI tool! https://lnkd.in/dabwaxqg