Skill UI

This SDK provides a comprehensive Python client for Azure AI Vision 4.0, enabling advanced image understanding tasks. It supports generating descriptive captions, extracting content tags, detecting objects and people with bounding boxes, performing OCR (text recognition), and suggesting smart crop regions. It is ideal for computer vision applications and detailed image content analysis.

Empowers building AI agents that perceive screens, plan via vision-language reasoning, and control mouse/keyboard for desktop automation, with attention to sandboxing and safety across Anthropic, OpenAI, and open-source paradigms.

Provides expert guidance on designing, implementing, and optimizing cutting-edge computer vision pipelines. Expertise spans real-time object detection (YOLO26), advanced promptable segmentation (SAM 3), and leveraging Vision Language Models (VLMs) for complex visual reasoning. Ideal for building high-performance systems in robotics, autonomous vehicles, and industrial inspection, including 3D reconstruction and edge deployment optimization.

Enables Claude to invoke the computer-vision-processor plugin, auto-generating Python calls to run object detection, classification, and segmentation on user-provided images for detailed analysis.

Comprehensive skill set covering the full lifecycle of advanced visual AI systems. Expertise includes object detection (YOLO, Faster R-CNN), instance and semantic segmentation (Mask R-CNN, SAM), model training using PyTorch, and production deployment optimization using ONNX and TensorRT for edge or cloud targets.

A comprehensive skill for leveraging massive pre-trained transformer models across diverse AI domains, including Natural Language Processing (NLP), Computer Vision (CV), and Audio. Use this skill for advanced tasks such as text generation, classification, question answering, summarization, object detection, and model fine-tuning on proprietary datasets. It provides streamlined pipelines and granular control for deployment.

A comprehensive technical module covering foundational deep learning principles and cutting-edge computer vision architectures, drawing from Yann LeCun's research. Topics include CNN fundamentals, backpropagation theory, and advanced self-supervised learning (SSL) models such as SimCLR, MAE, and the modern Joint Embedding Predictive Architecture (JEPA). Provides detailed theoretical explanations and complete PyTorch code for implementation.

Transformers.js allows developers to run state-of-the-art machine learning models directly within JavaScript/TypeScript environments (browser, Node.js). It supports a comprehensive range of tasks including NLP (text generation, summarization, translation), Computer Vision (image classification, object detection), and Audio processing. This capability enables building fully client-side, multimodal AI applications without requiring a dedicated Python backend or server infrastructure.

The Hugging Face Transformers library is a comprehensive toolkit providing access to thousands of pre-trained models across diverse domains including Natural Language Processing (NLP), Computer Vision, Audio, and Multimodal tasks. It enables developers to easily load models, perform high-level inference (e.g., text generation, classification, QA), and fine-tune models on custom datasets using streamlined APIs like Pipelines and the Trainer class. Ideal for building advanced, state-of-the-art AI applications.

This comprehensive skill manages the entire computer vision pipeline, from custom dataset annotation to model deployment. Users can fine-tune YOLO models on proprietary COCO-formatted datasets using an agent-driven workflow. Key features include hardware-aware training (supporting CUDA, MPS, ROCm), automatic export to optimized formats (TensorRT, CoreML, OpenVINO, ONNX), and seamless, one-click deployment as an active detection skill.

A comprehensive skill that automates the entire process of creating professional, production-ready thumbnails for articles or newsletters. It reads the article copy, proposes multiple composition concepts (e.g., bold claim, conceptual object), writes tailored prompts, calls the Gemini API for image generation, and evaluates the results using computer vision, returning ranked candidates with rationale.

This comprehensive guide helps computer vision researchers strategically position their work for the Asian Conference on Computer Vision (ACCV). It covers how to evaluate manuscript fit, restructure the narrative from general to conference-specific, optimize evidence presentation (including ablations and failure modes), and understand ACCV's regional niche compared to global flagships like CVPR or ECCV.