Virtual Home Staging: Inverse Rendering and Editing an Indoor Panorama under Natural Illumination

TL;DR

The paper tackles realistic 360° virtual home staging by enabling furniture editing in indoor panoramas under natural outdoor lighting. It combines calibrated indoor-outdoor HDR capture, panoramic furniture removal, automatic floor layout, and full 3D rendering to achieve photorealistic relighting and object insertion. Key contributions include the Cali-HDR calibrated HDR dataset, a panorama-aware furniture removal and inpainting pipeline, a rule-based automatic floor layout method, and a complete indoor virtual staging workflow. The approach enables practical, high-fidelity virtual staging for real-world indoor scenes, supporting applications in real estate and interior design through accurate global illumination and geometry-aware rendering.

Abstract

We propose a novel inverse rendering method that enables the transformation of existing indoor panoramas with new indoor furniture layouts under natural illumination. To achieve this, we captured indoor HDR panoramas along with real-time outdoor hemispherical HDR photographs. Indoor and outdoor HDR images were linearly calibrated with measured absolute luminance values for accurate scene relighting. Our method consists of three key components: (1) panoramic furniture detection and removal, (2) automatic floor layout design, and (3) global rendering with scene geometry, new furniture objects, and a real-time outdoor photograph. We demonstrate the effectiveness of our workflow in rendering indoor scenes under different outdoor illumination conditions. Additionally, we contribute a new calibrated HDR (Cali-HDR) dataset that consists of 137 calibrated indoor panoramas and their associated outdoor photographs.

Figures (8)

• Figure 1: Illustration of Proposed Rendering Pipeline: The captured scene (a) is filled with furniture objects. Detected furniture objects (b) are removed from the scene, and an empty scene (c) is restored. (d) New furniture objects fu20213d are inserted and rendered with real-time outdoor illumination.
• Figure 2: Our rendering pipeline consists of four modules: Indoor-Outdoor HDR Calibration (Sec. 3) calibrates the captured indoor and outdoor HDR photographs with measured absolute luminance values. Furniture Detection and Removal (Sec. 4) identifies and removes the target furniture objects from the scene. Automatic Floor Layout (Sec. 5) allows the automatic placement of multiple furniture objects. Indoor Virtual Staging (Sec. 6) achieves high-quality indoor virtual staging for furnished and empty scenes.
• Figure 3: Calibration Process: (a) Photographs from two cameras. (b) Cropped target regions. (c) Original luminance maps. (d) Luminance maps after HDR image captured by fisheye lens is scaled with $k_2$.
• Figure 4: Panoramic Scene Segmentation: (a) A single panorama is segmented into a set of 2D perspective images, and target furniture objects are detected. (b) The stitched panorama is processed to display furniture contours, and the rendered floor boundary is utilized to filter out solid contours that are not attached to the floor area. (c) Estimated tripod location zhi2022semantically, direct sunlight region, and the detected furniture areas are combined as the target mask.
• Figure 5: Comparison of Image Inpainting Methods: The target mask (from Fig. \ref{['fig_pano_furn_det']}(c)) is paired with input panorama (a) to remove the target region using PanoDR gkitsas2021panodr(b), LaMa suvorov2022resolution(c), and our method (d), respectively.