This tutorial extends a LangChain 1.0 multimodal RAG project by adding end‑to‑end image analysis and audio transcription features using Qwen3‑Omni, detailing data structures, utility classes, API changes, and Postman testing procedures.
Learn to leverage AI-powered multimodal extraction to identify material types in images, with step-by-step guidance on deploying cloud services, configuring storage, running visual models, and extracting structured attributes—empowering e‑commerce, design, and heritage applications.
This article demonstrates a practical OSINT workflow that combines social‑engineering clues, railway route research, terrain assessment, and feature‑based image analysis to pinpoint the exact shooting spot of a travel photograph taken on a train in northern China.
On January 17, the China Academy of Information and Communications Technology announced that Ant Security's self‑developed Tianjian multimodal content risk control system achieved the highest five‑star rating in both text and image assessments of the 2022 content review service evaluation, highlighting its advanced AI‑driven moderation capabilities.
This article explains how to programmatically parse container images using Google's go‑containerregistry library, covering basic concepts such as ImageIndex, Image Manifest, layers, and diffIDs, and demonstrates retrieving image metadata, system packages, and Java application dependencies through code examples and practical Go snippets.
Baidu’s Wànxiàng system processes billions of images and videos daily by extracting low‑ and high‑level features, linking related media, and aggregating semantic attributes in a scalable, timely architecture that leverages thousands of CPU, GPU, and FPGA cores to power accurate, low‑latency rich‑media search and recommendation.
The Xianyu team’s new “black‑tech” system automatically transforms UI mockup images into production‑ready code by detecting components with deep‑learning models, extracting layouts via OpenCV, and employing a modular stream‑oriented architecture of units, tasks, and server layers that enables rapid testing, flexible composition, and future enhancements such as improved container recognition and semantic understanding.
The paper presents an integrated multimodal quality assessment system for short‑video platforms that evaluates cover images, video content, and accompanying text using deep‑learning and handcrafted features—combining ResNet‑50, NetVLAD, TSN, VGGish, and XGBoost—to improve user experience, recommendation accuracy, and operational efficiency, with plans for optimization and modular deployment.
This presentation details how Beike Rental leverages big data and machine‑learning techniques to detect non‑authentic listings, build a four‑criterion true‑listing model, develop pricing and image‑analysis models, and design a merchant credit scoring system that improves service quality and market efficiency.
The talk by Tencent Cloud senior researcher Ji Yongnan reviews imaging modalities, low‑ to high‑level processing, the rise of deep‑learning CNNs over traditional handcrafted methods, and their deployment in classification, detection, segmentation, moderation, medical analysis, OCR, and other real‑world applications.
