-
Android 4.0+ support
-
ARMv7 and x86 based devices
-
Get the Caffe model and push it to Phone SDCard. For object detection, network(*.prototxt) should use ROILayer, you can refer to Fast-RCNN. For scene recognition(object recognition), it can use any caffe network and weight with memory input layer.
-
Build with Gradle. You can use Android studio to build
-
Object detection - Region-based Convolutional Networks detection Selective Search on Android + FastRCNN
-
Scene recognition - Convolutional neural networks trained on Places Input a picture of a place or scene and predicts it.
- Download and push the neccessary file to your phone. There should be model and weight in /sdcard/fastrcnn and /sdcard/vision_scene.
$ ./setup.sh
- Build and run the application using gradlew or you can open AndroidStudio to import this project
$ ./gradlew assembleDebug
$ adb install -r ./app/build/outputs/apk
Besides, you can change deep learning's model, weight, etc in VisionClassifierCreator.java
public class VisionClassifierCreator {
private final static String SCENE_MODEL_PATH = "..";
private final static String SCENE_WIEGHTS_PATH = "..";
private final static String SCENE_MEAN_FILE = "..";
private final static String SCENE_SYNSET_FILE = "..";
private final static String DETECT_MODEL_PATH = "..";
private final static String DETECT_WIEGHTS_PATH = "..";
private final static String DETECT_MEAN_FILE = "..";
private final static String DETECT_SYNSET_FILE = "..";
}- Send pull request
Copyright (C) 2015-2016 TzuTaLin
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.