Skill UI

Utilize the Kling AI API to generate high-quality, cinematic videos directly from descriptive text prompts. This advanced tool supports professional modes, camera movement control (pan, tilt, zoom), negative prompting, and synchronized native audio, making it ideal for professionals building robust text-to-video pipelines and creative content generation.

This comprehensive guide helps developers safely migrate between different versions of the Kling AI video generation model (e.g., v1.x to v2.x). It details breaking changes, parameter differences, feature availability matrices (including native audio and camera control), and provides code examples for parallel A/B testing and robust rollback strategies, ensuring seamless API adoption.

A specialized skill for generating hyper-realistic, humanized images using Google AI Studio (Gemini). It simulates the imperfections of real photography—such as natural lighting, slight camera shake, and depth of field—to create content that looks genuinely captured by a smartphone. Ideal for creating high-quality assets for social media (influencer style) or educational materials.

This skill provides real-time monocular depth estimation using advanced models like Depth Anything v2. It transforms live camera feeds by generating colorized depth maps, enabling depth overlays and detailed scene understanding. Crucially, it supports a 'privacy mode' that anonymizes visual identities by using depth-only transformation, preserving only the spatial layout and movement patterns for secure monitoring purposes. Optimized for both CoreML on macOS and PyTorch on other platforms.

This bridge provides bidirectional integration between Home Assistant (HA) and advanced AI pipelines. It streams HA camera feeds into the AI system for sophisticated analysis (detection, VLM, REID). Conversely, the analysis results—such as detected objects, descriptions, and person identities—are sent back to HA as image_processing entities, enabling powerful automation and smart notifications.

This benchmark evaluates Vision-Language Models (VLMs) on detecting potential safety hazards within indoor environments. It uses static camera frames, simulating real-world fixed security camera monitoring. The test covers 40 scenarios across 5 critical categories, including fire/smoke, electrical risks, trips/falls, child safety, and falling objects, providing a rigorous measure of VLM's practical safety inspection capability.

This skill provides high-performance, real-time object detection utilizing the Google Coral Edge TPU accelerator. It processes live camera frames to identify up to 80 COCO classes (e.g., person, car, dog). Optimized for native macOS and Linux deployment, it achieves extremely low latency (e.g., ~4ms at 320x320), making it ideal for embedded vision systems, surveillance, and low-latency monitoring applications.

This skill provides real-time object detection capabilities by leveraging the Google Coral Edge TPU accelerator. It operates natively within Windows Subsystem for Linux (WSL), allowing high-speed, low-latency inference on live camera frames. By utilizing hardware acceleration, it achieves impressive performance (e.g., ~4ms inference at 320x320), making it ideal for edge computing projects requiring rapid visual analysis and resource efficiency.

A comprehensive guide detailing the entire model lifecycle for RuView. It supports multiple advanced tracking methods including camera-free pose estimation (WiFlow), camera-supervised methods using MediaPipe, robust contrastive embeddings (RuVector), and domain generalization (MERIDIAN). Covers training, evaluation, and deployment across local GPU and cloud platforms.

This comprehensive guide assists researchers in converting an accepted, anonymous draft for the Conference on Learning Theory (COLT) into the final, polished camera-ready version for publication in PMLR proceedings. It covers critical steps such as de-anonymization (restoring author details, acknowledgements), consolidating all promised revisions into the main text, verifying PMLR metadata and BibTeX consistency, and preparing the full archival PDF.

A comprehensive guide detailing the submission rules and best practices for supplementary materials, including videos, appendices, code, and data, for top-tier robotics conferences. It outlines distinct submission channels, strict technical requirements (e.g., 250MB video cap, anonymity), and the necessary transition steps from the review phase to the final camera-ready publication format.

A comprehensive project management skill designed for authors preparing papers for the EACL conference. It maps out the entire single-cycle submission process, covering backward planning from the ARR deadline, rebuttal strategies, commitment decisions, and final camera-ready stages. Essential for navigating highly structured academic publication timelines and minimizing risk.