IndoorPanoDepth

We present a novel neural representation based method for depth estimation from a few panoramic images of different views. This is the official repo for the implementation of Depth Estimation from Indoor Panoramas with Neural Scene Representation (CVPR'2023).

Usage

For the Matterport3D and Stanford2D3D datasets, we adopt the rerendered version from 3D60. The brightness-adjusted dataset could be generated with 'adjust.py'.

Matterport3D

Copy the scenes to ./data/Matterport3D

SCENES="0_0b217f59904d4bdf85d35da2cab963471 1_0b724f78b3c04feeb3e744945517073d1 0_a2577698031844e7a5982c8ee0fecdeb1 0_9f2deaf4cf954d7aa43ce5dc70e7abbe1 0_7812e14df5e746388ff6cfe8b043950a1 4_0b724f78b3c04feeb3e744945517073d1 2_0b217f59904d4bdf85d35da2cab963471 1_7812e14df5e746388ff6cfe8b043950a1 47_a2577698031844e7a5982c8ee0fecdeb1 45_a2577698031844e7a5982c8ee0fecdeb1"
for scene in $SCENES; do
cp 3D60/Matterport3D/"$scene"_* ./data/Matterport3D
done

Direct run the following command.

sh Matterport3D.sh

---

Stanford2D3D

Copy the scenes to ./data/Stanford2D3D

SCENES="1_area_5a1 1_area_5b1 5_area_5a1 10_area_61 207_area_41"
for scene in $SCENES; do
cp 3D60/Stanford2D3D/"$scene"_* ./data/Stanford2D3D
done

Direct run the following command.

sh Stanford2D3D.sh

---

Our dataset

First download our dataset from https://1drv.ms/u/s!AmmYGRQ4ky-T1N0Bv8x7Oq_qQiKmNg?e=xUlxHR. Then, unzip it to the './data' fold as follows:

|-- code
    |-- data      
        |-- Matterport3D
        |-- Stanford2D3D
        |-- ours 
            |-- bedroom  
            ...

Finally, run the command

sh ours.sh

Acknowledgement

The main framework is borrowed from NeuS. The 3D models used for rendering dataset are from Flavio, Della, Tommasa, Christophe, Seux and Tadeusz. Thanks for these great works.