The repository contains an Android project which uses the MiDaS model to perform monocular depth estimation. You can find the official Android example here -> https://github.com/isl-org/MiDaS/tree/master/mobile/android
This project uses the TFLite model from the MiDaS's TensorFlow Hub repo. The following features are included in the project,
- Well documented code with links to SO answers wherever required.
- Uses latest APIs like CameraX and Kotlin Coroutines.
- No use of heavy packages like OpenCV to process and display the depth map. The application is coded in a Android friendly manner.
- Works for both front and rear camera and also in portrait and landscape orientations.
compileSdk 30
applicationId "com.shubham0204.ml.depthestimation"
minSdk 23
targetSdk 30
androidGradlePluginVersion 7.0.0
gradlePluginVersion 7.0.2
@article{DBLP:journals/corr/abs-1907-01341,
author = {Katrin Lasinger and
Ren{\'{e}} Ranftl and
Konrad Schindler and
Vladlen Koltun},
title = {Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-Shot
Cross-Dataset Transfer},
journal = {CoRR},
volume = {abs/1907.01341},
year = {2019},
url = {http://arxiv.org/abs/1907.01341},
archivePrefix = {arXiv},
eprint = {1907.01341},
timestamp = {Mon, 08 Jul 2019 14:12:33 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-1907-01341.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
Copyright 2021 Shubham Panchal
Licensed under the Apache License, Version 2.0 (the "License");
You may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.