Tagged articles
depth estimation
16 articles · Page 1 of 1
Xiaomi's WorldCast framework enables generating 3D camera movement videos from a single photo by combining depth-trajectory scale alignment, geometry-aligned memory attention, and a new CLACT benchmark, achieving state-of-the-art camera control and identity consistency on REDMI K100.
3D camera movementCLACT datasetWorldCast
0 likes · 10 min read

The article explains how the Chinese Ministry directives push robots from demo to operational mode, introduces LingBot‑Vision’s boundary‑centric masked modeling that embeds spatial structure into pre‑training, and shows LingBot‑Depth 2.0’s masked depth modeling dramatically improve dense depth perception on challenging surfaces while using far less data than prior SOTA models.
Boundary-Centric Masked ModelingLingBot-VisionSpace-Native Vision
0 likes · 15 min read

Meta’s VLM³ demonstrates that a plain vision‑language model, when trained on large‑scale data with simple camera‑focal‑length and pixel‑space normalization, matches or surpasses expert 3D vision models across monocular depth estimation, object‑level understanding, pixel‑matching and camera‑pose tasks, eliminating the need for task‑specific architectures, loss functions, data augmentations or regression formulations.
3D visionMetaVLM
0 likes · 6 min read

Meta and Princeton introduce VLM³, a unified vision‑language framework built on Qwen3‑VL‑4B that models depth estimation, object‑level 3D understanding, pixel matching and camera pose estimation without extra encoders, achieving up to 0.90 depth accuracy and outperforming larger specialist models on multiple benchmarks.
3D PerceptionQwen3-VL-4BVision-Language Models
0 likes · 15 min read

The ByteDance‑Seed team introduces Depth Anything 3 (DA3), a minimalist visual‑geometry model that uses a vanilla Transformer backbone and depth‑ray representation to jointly predict depth and camera pose from any number of images, achieving state‑of‑the‑art performance with a 35.7% gain in pose accuracy and a 23.6% improvement in geometric precision over prior methods.
3D visionDA3Transformer
0 likes · 6 min read

Leveraging 3D vision and AIGC, JD Retail’s R&D team converts abundant 2D video assets into high‑quality stereoscopic 3D space videos through a pipeline that includes monocular depth estimation, novel view synthesis, multi‑branch inpainting, and MV‑HEVC encoding, validated by ICME 2025 and a new StereoV1K dataset.
3D videoAIGCComputer Vision
0 likes · 26 min read


JD TechApr 21, 2025 · Artificial Intelligence
This article presents a comprehensive AI‑driven pipeline that converts 2D video into immersive 3D spatial video by leveraging monocular depth estimation, depth‑warping novel view synthesis, a multi‑branch inpainting module, a large‑scale StereoV1K dataset, and efficient MV‑HEVC compression, with results validated at ICME 2025 and deployed in JD Vision services.
3D videoAIAIGC
0 likes · 20 min read

This work presents an end‑to‑end AI pipeline that transforms existing monocular 2D videos into immersive 3D spatial streams by combining DINO‑v2‑based depth estimation, multi‑branch view synthesis, and MV‑HEVC encoding, achieving up to 33 % BD‑Rate reduction, 31 % speed gains, state‑of‑the‑art visual quality, and real‑time production suitability, validated on the new StereoV1K benchmark and deployed in JD.Vision’s e‑commerce catalog.
3D videoAI GenerationAIGC
0 likes · 21 min read


AIWalkerJan 22, 2024 · Artificial Intelligence
Depth Anything introduces a highly practical monocular depth estimation model that leverages a 62‑million‑image unlabeled dataset, teacher‑student training, strong data perturbations, and DINOv2‑based semantic supervision to achieve zero‑shot capability and state‑of‑the‑art performance over MiDaS across multiple benchmarks.
Computer VisionDINOv2Monocular Depth
0 likes · 8 min read

This article examines the challenges of 3D vision data acquisition and explains how Baidu's PaddleDepth and Paddle3D toolkits provide low‑cost depth collection, super‑resolution, and end‑to‑end perception pipelines, showcasing performance on KITTI and Middlebury datasets with code examples.
3D visionComputer VisionPaddle3D
0 likes · 12 min read

This article explains the optical principles of bokeh (scatter blur), describes a depth‑aware variable‑focus algorithm developed by Kuaishou’s audio‑video team, and details practical optimizations such as saliency detection, edge‑preserving weighting, and adaptive spot‑light effects that enable realistic, customizable holiday video filters.
BokehComputer VisionVideo Effects
0 likes · 11 min read

The paper introduces D4, a self‑augmented unpaired image dehazing framework that decomposes the transmission map into fog density and scene depth, enabling realistic fog synthesis for data augmentation and achieving superior dehazing performance with fewer parameters and FLOPs on multiple benchmarks.
CVPR2022Computer Visiondensity decomposition
0 likes · 14 min read

This article presents Kuaishou's mobile mixed reality (MR) system, detailing its integration of deep learning, SLAM, and scene reconstruction for real‑time spatial computing, the design of a monocular depth‑estimation model, a lightweight 3D rendering engine, and its deployment across iOS and Android devices with various user‑facing effects.
Computer VisionKuaishouMobile AR
0 likes · 16 min read

The paper presents HDES‑Net, a lightweight real‑time monocular human depth estimation and segmentation network designed for embedded platforms, using MobileNetV1 backbone with ASPP and depth‑wise separable convolutions, achieving high accuracy on CAD‑60 and EPFL‑RGBD datasets while running at up to 199.93 FPS on a Tesla P40 and 17.23 FPS on a Jetson Nano after TensorRT optimization.
Embedded AIHDES-Netdepth estimation
0 likes · 8 min read

This article explains the fundamentals of virtual reality and its distinction from augmented reality, describes VR hardware, outlines depth‑estimation and eye‑tracking methods such as projection, Hough transform, AdaBoost and sample matching, discusses Sobel edge detection, and explores the importance of audio, haptic feedback, and immersive VR applications in education.
ARComputer VisionImmersive Education
0 likes · 11 min read

iQIYI’s deep‑learning pipeline converts single‑view images into high‑quality stereo pairs for VR by training on side‑by‑side 3D movies, employing a Monodepth‑based encoder‑decoder, a CVAE to encode camera parameters, ConvLSTM for temporal consistency, and disparity‑guided inpainting to fill occlusion holes, achieving stable, continuous depth maps validated through extensive human 3‑D effect assessments.
2D-to-3DVRdepth estimation
0 likes · 12 min read
