I had experimented with the camera tracking posture using Apple MLKit (all local) and showing an onscreen alert. Using AirPods seems way smarter. Is the underlying API public?

Such a great plug-and-play React Native library that exposes MLKit by Luiz Carlos. It integrates seamlessly with Vision Camera v5 SkiaCamera. And here I am stress testing it 😅.

That’s impressive, a full facial emotion is more comprehensive, what was your base model for detection and landmarks. This one actually uses MLKit face detection from google, and it has Smile probability which is based on eyes and mostly mouth corners.

bet they just had one of their in house developer daycare twats collect 4k/wk for 6 months while adding the summarizer mlkit to a chrome extension between the free cliff bars and Red Bull; and that guy probably had some contractor do it in a weekend for 50 bucks

Does that mean they're going to stop offering the summarizer MLkit as a downloadable package?

AI流行ってるからツール全部AIが入ってると思う人がいるってことなのかな🤔 ちょっと前に画像から文字起こしするツールを試しに作ってみましたが、AIのAPIに投げるとめっちゃお金かかるから、結局MLkitで色んな工夫をして低コストに😵

react-native-mlkit-ocr あんまりメンテナンスされてないみたいで他の最新環境を要求するライブラリと競合してビルドエラーが発生するみたい

Coming up with a banger library with help of @margelo_com and @mrousavy's Vision Camera V5, Skia, Vision API from iOS and MLKit from Android. Launching soon. Special thanks to @RiteshRk14 for helping out on picking Vision & MLKit over Media pose.

🩳 KIT DA BERMA MAIS VENDIDA DO MLKit com 3 Bermudas Voker Lisas Treino Academia ❌ De: 117 ✅ Por: 62 🎟️ Use o Cupom: MODAML 👉 meli.la/2nsT2Y9

🆕 【MLKit x Gemini Nano 4】オンデバイス生成AIでマルチモーダルな推論をしてみる ✅Android端末内でマルチモーダル推論を実現 ✅ML Kitを使った最新モデルの実装手順を詳しく解説 ぜひご一読ください！ qiita.com/FrozenPudding/item… #Qiita #Qiita_jtech #LLM #AI

try now: npx expo install expo-mlkit-ocr npmjs.com/package/expo-mlkit…

today I built a react native / expo library called @expo -mlkit-ocr. on-device OCR from images, fast and no cloud needed. includes OCRTextOverlay, everything 1-line setup, fully customizable, works with any Image component, and supports multi-text selection. plug & play 🫡

We're looking for a Research Assistant (Embedded AI Systems) at Language Technologies Research Center, IIITH. This role is focused on building and optimizing real-world on-device AI systems. The work involves developing embedded Linux pipelines (Buildroot-based devices), handling real-time audio/video data transfer, and running AI models directly on edge hardware. On the mobile side, you'll work with on-device models (Gemma with LiteRT, MLKit, speech systems), and upcoming work includes optimizing inference using Qualcomm's QNN stack on NPUs/GPUs. This is hands-on systems work. You should be interested in low-level programming (C/C ), comfortable exploring Linux/embedded systems, and curious about performance optimization and hardware-accelerated AI. Prior experience helps, but willingness to learn and build matters more. The work is tied to building AI-enabled accessibility systems for visually and speech-impaired users, so there is a clear real-world impact alongside the technical depth. Duration is 2 months (on-campus), with possible extension. Stipend is ₹10,000/month. If this aligns with what you want to work on, reach out.

Stop re-computing the same context 1,000 times. Read more about this -> linkedin.com/feed/update/urn… hashtag#AndroidDev hashtag#GenerativeAI hashtag#MLKit hashtag#GeminiNano hashtag#OnDeviceAI

VisionCamera V4's Barcode Scanner used Apple's on iOS and MLKit on Android. Different formats, different bugs. V5 now uses MLKit on both platforms. Consistent everywhere. Additionally, there's a native Object Output (Apple-only)