HSCNN+: Advanced CNN-Based Hyperspectral Recovery from RGB Images in CVPRW 2018 (Winner of NTIRE Challenge)

RGB-based Category-level Object Pose Estimation via Decoupled Metric Scale Recovery

This is the rep of 'End-to-end Hand Mesh Recovery from a Monocular RGB Image'

This is the C implementation of the paper "Color-guided Depth Recovery from RGB-D Data Using an Adaptive Auto-regressive Model" in TIP 2014. It is the same as the MATLAB source code provided by the author. However, it can run much faster and can handle much larger depth maps.

This is the rep of 'End-to-end Hand Mesh Recovery from a Monocular RGB Image'

Official repository for PURA: Parameter Update-Recovery Test-Time Adaption for RGB-T Tracking [CVPR 2025]

The code of paper "RGB-Guided Depth Map Recovery by Two-Stage Coarse-to-fine Dense CRF Models "

A project on recovery and reconstruction of the 3D Human Mesh of an individual from a single RGB image, which will be expanded into a video, and possibly live video. Implemented in PyTorch, and based on the research paper titled, "End-to-end Recovery of Human Shape and Pose."

OPENCV project， recover Sergei Prokudin-Gorsky 's piucture sets

No description available

2D pose estimation and human mesh recovery from rgb image.

Official page of the Human Mesh Modeling and Recovery from RGB and LiDAR

Official page of the Human Mesh Modeling and Recovery from RGB and LiDAR Tutorial

A recovery enhanced category specific RGB Image Steganography using Yolo v8 and Cross Axis MLP

A technical article on reformulating RGB-D cross-sensor mapping as a sparse geometric query for lightweight 3D recovery.

This repository implements a deep learning approach for spectral super-resolution, transforming 3-channel RGB images into 31-channel hyperspectral images. The model is trained on the NTIRE 2020 Spectral Recovery Challenge dataset.

A comprehensive steganography system that securely embeds compressed data within color images using Huffman coding and histogram shifting algorithms. Features multi-channel RGB encoding with 3x capacity increase, lossless data recovery, and comprehensive quality analysis tools.

This project implements a robust pipeline for 3D hand pose estimation from single RGB images. We leverage the state-of-the-art HaMeR (Hand Mesh Recovery) model, integrated as a Git submodule to ensure maintainability and reproducibility.

This project implements image encryption and decryption using pixel manipulation in Python. Each RGB pixel value is modified using a secret key with modulo arithmetic. The program ensures original image recovery after decryption and demonstrates basic image processing and security concepts.