Project import generated by Copybara.
GitOrigin-RevId: f7d09ed033907b893638a8eb4148efa11c0f09a6
This commit is contained in:
MediaPipe Team
+3
-4
@@ -36,9 +36,8 @@ android_binary(
|
||||
name = "facedetectioncpu",
|
||||
srcs = glob(["*.java"]),
|
||||
assets = [
|
||||
"//mediapipe/graphs/face_detection:mobile_cpu.binarypb",
|
||||
"//mediapipe/models:face_detection_front.tflite",
|
||||
"//mediapipe/models:face_detection_front_labelmap.txt",
|
||||
"//mediapipe/graphs/face_detection:face_detection_mobile_cpu.binarypb",
|
||||
"//mediapipe/modules/face_detection:face_detection_front.tflite",
|
||||
],
|
||||
assets_dir = "",
|
||||
manifest = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:AndroidManifest.xml",
|
||||
@@ -47,7 +46,7 @@ android_binary(
|
||||
"appName": "Face Detection (CPU)",
|
||||
"mainActivity": "com.google.mediapipe.apps.basic.MainActivity",
|
||||
"cameraFacingFront": "True",
|
||||
"binaryGraphName": "mobile_cpu.binarypb",
|
||||
"binaryGraphName": "face_detection_mobile_cpu.binarypb",
|
||||
"inputVideoStreamName": "input_video",
|
||||
"outputVideoStreamName": "output_video",
|
||||
"flipFramesVertically": "True",
|
||||
|
||||
+3
-4
@@ -36,9 +36,8 @@ android_binary(
|
||||
name = "facedetectiongpu",
|
||||
srcs = glob(["*.java"]),
|
||||
assets = [
|
||||
"//mediapipe/graphs/face_detection:mobile_gpu.binarypb",
|
||||
"//mediapipe/models:face_detection_front.tflite",
|
||||
"//mediapipe/models:face_detection_front_labelmap.txt",
|
||||
"//mediapipe/graphs/face_detection:face_detection_mobile_gpu.binarypb",
|
||||
"//mediapipe/modules/face_detection:face_detection_front.tflite",
|
||||
],
|
||||
assets_dir = "",
|
||||
manifest = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:AndroidManifest.xml",
|
||||
@@ -47,7 +46,7 @@ android_binary(
|
||||
"appName": "Face Detection",
|
||||
"mainActivity": "com.google.mediapipe.apps.basic.MainActivity",
|
||||
"cameraFacingFront": "True",
|
||||
"binaryGraphName": "mobile_gpu.binarypb",
|
||||
"binaryGraphName": "face_detection_mobile_gpu.binarypb",
|
||||
"inputVideoStreamName": "input_video",
|
||||
"outputVideoStreamName": "output_video",
|
||||
"flipFramesVertically": "True",
|
||||
|
||||
+1
-2
@@ -37,8 +37,7 @@ android_binary(
|
||||
srcs = glob(["*.java"]),
|
||||
assets = [
|
||||
"//mediapipe/graphs/hand_tracking:hand_detection_mobile_gpu.binarypb",
|
||||
"//mediapipe/models:palm_detection.tflite",
|
||||
"//mediapipe/models:palm_detection_labelmap.txt",
|
||||
"//mediapipe/modules/palm_detection:palm_detection.tflite",
|
||||
],
|
||||
assets_dir = "",
|
||||
manifest = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:AndroidManifest.xml",
|
||||
|
||||
@@ -37,10 +37,9 @@ android_binary(
|
||||
srcs = glob(["*.java"]),
|
||||
assets = [
|
||||
"//mediapipe/graphs/hand_tracking:hand_tracking_mobile_gpu.binarypb",
|
||||
"//mediapipe/models:handedness.txt",
|
||||
"//mediapipe/models:hand_landmark.tflite",
|
||||
"//mediapipe/models:palm_detection.tflite",
|
||||
"//mediapipe/models:palm_detection_labelmap.txt",
|
||||
"//mediapipe/modules/hand_landmark:handedness.txt",
|
||||
"//mediapipe/modules/hand_landmark:hand_landmark.tflite",
|
||||
"//mediapipe/modules/palm_detection:palm_detection.tflite",
|
||||
],
|
||||
assets_dir = "",
|
||||
manifest = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:AndroidManifest.xml",
|
||||
|
||||
+45
-46
@@ -18,76 +18,75 @@ import android.os.Bundle;
|
||||
import android.util.Log;
|
||||
import com.google.mediapipe.formats.proto.LandmarkProto.NormalizedLandmark;
|
||||
import com.google.mediapipe.formats.proto.LandmarkProto.NormalizedLandmarkList;
|
||||
import com.google.mediapipe.framework.AndroidPacketCreator;
|
||||
import com.google.mediapipe.framework.Packet;
|
||||
import com.google.mediapipe.framework.PacketGetter;
|
||||
import com.google.protobuf.InvalidProtocolBufferException;
|
||||
import java.util.HashMap;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
|
||||
/** Main activity of MediaPipe hand tracking app. */
|
||||
public class MainActivity extends com.google.mediapipe.apps.basic.MainActivity {
|
||||
private static final String TAG = "MainActivity";
|
||||
|
||||
private static final String OUTPUT_HAND_PRESENCE_STREAM_NAME = "hand_presence";
|
||||
private static final String INPUT_NUM_HANDS_SIDE_PACKET_NAME = "num_hands";
|
||||
private static final String OUTPUT_LANDMARKS_STREAM_NAME = "hand_landmarks";
|
||||
// Max number of hands to detect/process.
|
||||
private static final int NUM_HANDS = 2;
|
||||
|
||||
@Override
|
||||
protected void onCreate(Bundle savedInstanceState) {
|
||||
super.onCreate(savedInstanceState);
|
||||
|
||||
processor.addPacketCallback(
|
||||
OUTPUT_HAND_PRESENCE_STREAM_NAME,
|
||||
(packet) -> {
|
||||
Boolean handPresence = PacketGetter.getBool(packet);
|
||||
if (!handPresence) {
|
||||
Log.d(
|
||||
TAG,
|
||||
"[TS:" + packet.getTimestamp() + "] Hand presence is false, no hands detected.");
|
||||
}
|
||||
});
|
||||
AndroidPacketCreator packetCreator = processor.getPacketCreator();
|
||||
Map<String, Packet> inputSidePackets = new HashMap<>();
|
||||
inputSidePackets.put(INPUT_NUM_HANDS_SIDE_PACKET_NAME, packetCreator.createInt32(NUM_HANDS));
|
||||
processor.setInputSidePackets(inputSidePackets);
|
||||
|
||||
// To show verbose logging, run:
|
||||
// adb shell setprop log.tag.MainActivity VERBOSE
|
||||
if (Log.isLoggable(TAG, Log.VERBOSE)) {
|
||||
processor.addPacketCallback(
|
||||
OUTPUT_LANDMARKS_STREAM_NAME,
|
||||
(packet) -> {
|
||||
byte[] landmarksRaw = PacketGetter.getProtoBytes(packet);
|
||||
try {
|
||||
NormalizedLandmarkList landmarks = NormalizedLandmarkList.parseFrom(landmarksRaw);
|
||||
if (landmarks == null) {
|
||||
Log.v(TAG, "[TS:" + packet.getTimestamp() + "] No hand landmarks.");
|
||||
return;
|
||||
}
|
||||
// Note: If hand_presence is false, these landmarks are useless.
|
||||
OUTPUT_LANDMARKS_STREAM_NAME,
|
||||
(packet) -> {
|
||||
Log.v(TAG, "Received multi-hand landmarks packet.");
|
||||
List<NormalizedLandmarkList> multiHandLandmarks =
|
||||
PacketGetter.getProtoVector(packet, NormalizedLandmarkList.parser());
|
||||
Log.v(
|
||||
TAG,
|
||||
"[TS:"
|
||||
+ packet.getTimestamp()
|
||||
+ "] #Landmarks for hand: "
|
||||
+ landmarks.getLandmarkCount());
|
||||
Log.v(TAG, getLandmarksDebugString(landmarks));
|
||||
} catch (InvalidProtocolBufferException e) {
|
||||
Log.e(TAG, "Couldn't Exception received - " + e);
|
||||
return;
|
||||
}
|
||||
});
|
||||
+ "] "
|
||||
+ getMultiHandLandmarksDebugString(multiHandLandmarks));
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
private static String getLandmarksDebugString(NormalizedLandmarkList landmarks) {
|
||||
int landmarkIndex = 0;
|
||||
String landmarksString = "";
|
||||
for (NormalizedLandmark landmark : landmarks.getLandmarkList()) {
|
||||
landmarksString +=
|
||||
"\t\tLandmark["
|
||||
+ landmarkIndex
|
||||
+ "]: ("
|
||||
+ landmark.getX()
|
||||
+ ", "
|
||||
+ landmark.getY()
|
||||
+ ", "
|
||||
+ landmark.getZ()
|
||||
+ ")\n";
|
||||
++landmarkIndex;
|
||||
private String getMultiHandLandmarksDebugString(List<NormalizedLandmarkList> multiHandLandmarks) {
|
||||
if (multiHandLandmarks.isEmpty()) {
|
||||
return "No hand landmarks";
|
||||
}
|
||||
return landmarksString;
|
||||
String multiHandLandmarksStr = "Number of hands detected: " + multiHandLandmarks.size() + "\n";
|
||||
int handIndex = 0;
|
||||
for (NormalizedLandmarkList landmarks : multiHandLandmarks) {
|
||||
multiHandLandmarksStr +=
|
||||
"\t#Hand landmarks for hand[" + handIndex + "]: " + landmarks.getLandmarkCount() + "\n";
|
||||
int landmarkIndex = 0;
|
||||
for (NormalizedLandmark landmark : landmarks.getLandmarkList()) {
|
||||
multiHandLandmarksStr +=
|
||||
"\t\tLandmark ["
|
||||
+ landmarkIndex
|
||||
+ "]: ("
|
||||
+ landmark.getX()
|
||||
+ ", "
|
||||
+ landmark.getY()
|
||||
+ ", "
|
||||
+ landmark.getZ()
|
||||
+ ")\n";
|
||||
++landmarkIndex;
|
||||
}
|
||||
++handIndex;
|
||||
}
|
||||
return multiHandLandmarksStr;
|
||||
}
|
||||
}
|
||||
|
||||
-64
@@ -1,64 +0,0 @@
|
||||
# Copyright 2019 The MediaPipe Authors.
|
||||
#
|
||||
# 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.
|
||||
|
||||
licenses(["notice"])
|
||||
|
||||
package(default_visibility = ["//visibility:private"])
|
||||
|
||||
cc_binary(
|
||||
name = "libmediapipe_jni.so",
|
||||
linkshared = 1,
|
||||
linkstatic = 1,
|
||||
deps = [
|
||||
"//mediapipe/graphs/hand_tracking:multi_hand_mobile_calculators",
|
||||
"//mediapipe/java/com/google/mediapipe/framework/jni:mediapipe_framework_jni",
|
||||
],
|
||||
)
|
||||
|
||||
cc_library(
|
||||
name = "mediapipe_jni_lib",
|
||||
srcs = [":libmediapipe_jni.so"],
|
||||
alwayslink = 1,
|
||||
)
|
||||
|
||||
android_binary(
|
||||
name = "multihandtrackinggpu",
|
||||
srcs = glob(["*.java"]),
|
||||
assets = [
|
||||
"//mediapipe/graphs/hand_tracking:multi_hand_tracking_mobile_gpu.binarypb",
|
||||
"//mediapipe/models:handedness.txt",
|
||||
"//mediapipe/models:hand_landmark.tflite",
|
||||
"//mediapipe/models:palm_detection.tflite",
|
||||
"//mediapipe/models:palm_detection_labelmap.txt",
|
||||
],
|
||||
assets_dir = "",
|
||||
manifest = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:AndroidManifest.xml",
|
||||
manifest_values = {
|
||||
"applicationId": "com.google.mediapipe.apps.multihandtrackinggpu",
|
||||
"appName": "Multi-hand Tracking",
|
||||
"mainActivity": ".MainActivity",
|
||||
"cameraFacingFront": "True",
|
||||
"binaryGraphName": "multi_hand_tracking_mobile_gpu.binarypb",
|
||||
"inputVideoStreamName": "input_video",
|
||||
"outputVideoStreamName": "output_video",
|
||||
"flipFramesVertically": "True",
|
||||
},
|
||||
multidex = "native",
|
||||
deps = [
|
||||
":mediapipe_jni_lib",
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:basic_lib",
|
||||
"//mediapipe/framework/formats:landmark_java_proto_lite",
|
||||
"//mediapipe/java/com/google/mediapipe/framework:android_framework",
|
||||
],
|
||||
)
|
||||
-80
@@ -1,80 +0,0 @@
|
||||
// Copyright 2019 The MediaPipe Authors.
|
||||
//
|
||||
// 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.
|
||||
|
||||
package com.google.mediapipe.apps.multihandtrackinggpu;
|
||||
|
||||
import android.os.Bundle;
|
||||
import android.util.Log;
|
||||
import com.google.mediapipe.formats.proto.LandmarkProto.NormalizedLandmark;
|
||||
import com.google.mediapipe.formats.proto.LandmarkProto.NormalizedLandmarkList;
|
||||
import com.google.mediapipe.framework.PacketGetter;
|
||||
import java.util.List;
|
||||
|
||||
/** Main activity of MediaPipe multi-hand tracking app. */
|
||||
public class MainActivity extends com.google.mediapipe.apps.basic.MainActivity {
|
||||
private static final String TAG = "MainActivity";
|
||||
|
||||
private static final String OUTPUT_LANDMARKS_STREAM_NAME = "multi_hand_landmarks";
|
||||
|
||||
@Override
|
||||
protected void onCreate(Bundle savedInstanceState) {
|
||||
super.onCreate(savedInstanceState);
|
||||
|
||||
// To show verbose logging, run:
|
||||
// adb shell setprop log.tag.MainActivity VERBOSE
|
||||
if (Log.isLoggable(TAG, Log.VERBOSE)) {
|
||||
processor.addPacketCallback(
|
||||
OUTPUT_LANDMARKS_STREAM_NAME,
|
||||
(packet) -> {
|
||||
Log.v(TAG, "Received multi-hand landmarks packet.");
|
||||
List<NormalizedLandmarkList> multiHandLandmarks =
|
||||
PacketGetter.getProtoVector(packet, NormalizedLandmarkList.parser());
|
||||
Log.v(
|
||||
TAG,
|
||||
"[TS:"
|
||||
+ packet.getTimestamp()
|
||||
+ "] "
|
||||
+ getMultiHandLandmarksDebugString(multiHandLandmarks));
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
private String getMultiHandLandmarksDebugString(List<NormalizedLandmarkList> multiHandLandmarks) {
|
||||
if (multiHandLandmarks.isEmpty()) {
|
||||
return "No hand landmarks";
|
||||
}
|
||||
String multiHandLandmarksStr = "Number of hands detected: " + multiHandLandmarks.size() + "\n";
|
||||
int handIndex = 0;
|
||||
for (NormalizedLandmarkList landmarks : multiHandLandmarks) {
|
||||
multiHandLandmarksStr +=
|
||||
"\t#Hand landmarks for hand[" + handIndex + "]: " + landmarks.getLandmarkCount() + "\n";
|
||||
int landmarkIndex = 0;
|
||||
for (NormalizedLandmark landmark : landmarks.getLandmarkList()) {
|
||||
multiHandLandmarksStr +=
|
||||
"\t\tLandmark ["
|
||||
+ landmarkIndex
|
||||
+ "]: ("
|
||||
+ landmark.getX()
|
||||
+ ", "
|
||||
+ landmark.getY()
|
||||
+ ", "
|
||||
+ landmark.getZ()
|
||||
+ ")\n";
|
||||
++landmarkIndex;
|
||||
}
|
||||
++handIndex;
|
||||
}
|
||||
return multiHandLandmarksStr;
|
||||
}
|
||||
}
|
||||
+127
-38
@@ -1,4 +1,4 @@
|
||||
# Copyright 2019 The MediaPipe Authors.
|
||||
# Copyright 2020 The MediaPipe Authors.
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
@@ -12,16 +12,64 @@
|
||||
# See the License for the specific language governing permissions and
|
||||
# limitations under the License.
|
||||
|
||||
load("@bazel_skylib//lib:selects.bzl", "selects")
|
||||
load(":build_defs.bzl", "generate_manifest_values")
|
||||
|
||||
licenses(["notice"])
|
||||
|
||||
package(default_visibility = ["//visibility:private"])
|
||||
|
||||
config_setting(
|
||||
name = "use_chair",
|
||||
define_values = {
|
||||
"chair": "true",
|
||||
},
|
||||
)
|
||||
|
||||
config_setting(
|
||||
name = "use_cup",
|
||||
define_values = {
|
||||
"cup": "true",
|
||||
},
|
||||
)
|
||||
|
||||
config_setting(
|
||||
name = "use_camera",
|
||||
define_values = {
|
||||
"camera": "true",
|
||||
},
|
||||
)
|
||||
|
||||
config_setting(
|
||||
name = "use_shoe_1stage",
|
||||
define_values = {
|
||||
"shoe_1stage": "true",
|
||||
},
|
||||
)
|
||||
|
||||
config_setting(
|
||||
name = "use_chair_1stage",
|
||||
define_values = {
|
||||
"chair_1stage": "true",
|
||||
},
|
||||
)
|
||||
|
||||
selects.config_setting_group(
|
||||
name = "1stage",
|
||||
match_any = [
|
||||
":use_shoe_1stage",
|
||||
":use_chair_1stage",
|
||||
],
|
||||
)
|
||||
|
||||
cc_binary(
|
||||
name = "libmediapipe_jni.so",
|
||||
linkshared = 1,
|
||||
linkstatic = 1,
|
||||
deps = [
|
||||
"//mediapipe/graphs/object_detection_3d:mobile_calculators",
|
||||
deps = select({
|
||||
"//conditions:default": ["//mediapipe/graphs/object_detection_3d:mobile_calculators"],
|
||||
":1stage": ["//mediapipe/graphs/object_detection_3d:mobile_calculators_1stage"],
|
||||
}) + [
|
||||
"//mediapipe/java/com/google/mediapipe/framework/jni:mediapipe_framework_jni",
|
||||
],
|
||||
)
|
||||
@@ -32,67 +80,108 @@ cc_library(
|
||||
alwayslink = 1,
|
||||
)
|
||||
|
||||
# To use the "chair" model instead of the default "shoes" model,
|
||||
# add "--define chair=true" to the bazel build command.
|
||||
config_setting(
|
||||
name = "use_chair_model",
|
||||
define_values = {
|
||||
"chair": "true",
|
||||
},
|
||||
)
|
||||
|
||||
genrule(
|
||||
name = "binary_graph",
|
||||
srcs = select({
|
||||
"//conditions:default": ["//mediapipe/graphs/object_detection_3d:mobile_gpu_binary_graph_shoe"],
|
||||
":use_chair_model": ["//mediapipe/graphs/object_detection_3d:mobile_gpu_binary_graph_chair"],
|
||||
"//conditions:default": ["//mediapipe/graphs/object_detection_3d:mobile_gpu_binary_graph"],
|
||||
":1stage": ["//mediapipe/graphs/object_detection_3d:mobile_gpu_1stage_binary_graph"],
|
||||
}),
|
||||
outs = ["object_detection_3d.binarypb"],
|
||||
cmd = "cp $< $@",
|
||||
)
|
||||
|
||||
MODELS_DIR = "//mediapipe/models"
|
||||
|
||||
genrule(
|
||||
name = "model",
|
||||
srcs = select({
|
||||
"//conditions:default": ["//mediapipe/models:object_detection_3d_sneakers.tflite"],
|
||||
":use_chair_model": ["//mediapipe/models:object_detection_3d_chair.tflite"],
|
||||
"//conditions:default": [MODELS_DIR + ":object_detection_3d_sneakers.tflite"],
|
||||
":use_chair": [MODELS_DIR + ":object_detection_3d_chair.tflite"],
|
||||
":use_cup": [MODELS_DIR + ":object_detection_3d_cup.tflite"],
|
||||
":use_camera": [MODELS_DIR + ":object_detection_3d_camera.tflite"],
|
||||
":use_shoe_1stage": [MODELS_DIR + ":object_detection_3d_sneakers_1stage.tflite"],
|
||||
":use_chair_1stage": [MODELS_DIR + ":object_detection_3d_chair_1stage.tflite"],
|
||||
}),
|
||||
outs = ["object_detection_3d.tflite"],
|
||||
cmd = "cp $< $@",
|
||||
)
|
||||
|
||||
MANIFESTS_DIR = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/manifests"
|
||||
|
||||
android_library(
|
||||
name = "manifest_lib",
|
||||
exports_manifest = 1,
|
||||
manifest = select({
|
||||
"//conditions:default": MANIFESTS_DIR + ":AndroidManifestSneaker.xml",
|
||||
":use_chair": MANIFESTS_DIR + ":AndroidManifestChair.xml",
|
||||
":use_cup": MANIFESTS_DIR + ":AndroidManifestCup.xml",
|
||||
":use_camera": MANIFESTS_DIR + ":AndroidManifestCamera.xml",
|
||||
":use_shoe_1stage": MANIFESTS_DIR + ":AndroidManifestSneaker.xml",
|
||||
":use_chair_1stage": MANIFESTS_DIR + ":AndroidManifestChair.xml",
|
||||
}),
|
||||
deps = [
|
||||
"//third_party:opencv",
|
||||
"@maven//:androidx_concurrent_concurrent_futures",
|
||||
"@maven//:com_google_guava_guava",
|
||||
],
|
||||
)
|
||||
|
||||
ASSETS_DIR = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/assets"
|
||||
|
||||
genrule(
|
||||
name = "mesh",
|
||||
srcs = select({
|
||||
"//conditions:default": [ASSETS_DIR + "/sneaker:model.obj.uuu"],
|
||||
":use_chair": [ASSETS_DIR + "/chair:model.obj.uuu"],
|
||||
":use_cup": [ASSETS_DIR + "/cup:model.obj.uuu"],
|
||||
":use_camera": [ASSETS_DIR + "/camera:model.obj.uuu"],
|
||||
":use_shoe_1stage": [ASSETS_DIR + "/sneaker:model.obj.uuu"],
|
||||
":use_chair_1stage": [ASSETS_DIR + "/chair:model.obj.uuu"],
|
||||
}),
|
||||
outs = ["model.obj.uuu"],
|
||||
cmd = "cp $< $@",
|
||||
)
|
||||
|
||||
genrule(
|
||||
name = "texture",
|
||||
srcs = select({
|
||||
"//conditions:default": [ASSETS_DIR + "/sneaker:texture.jpg"],
|
||||
":use_chair": [ASSETS_DIR + "/chair:texture.jpg"],
|
||||
":use_cup": [ASSETS_DIR + "/cup:texture.jpg"],
|
||||
":use_camera": [ASSETS_DIR + "/camera:texture.jpg"],
|
||||
":use_shoe_1stage": [ASSETS_DIR + "/sneaker:texture.jpg"],
|
||||
":use_chair_1stage": [ASSETS_DIR + "/chair:texture.jpg"],
|
||||
}),
|
||||
outs = ["texture.jpg"],
|
||||
cmd = "cp $< $@",
|
||||
)
|
||||
|
||||
android_binary(
|
||||
name = "objectdetection3d",
|
||||
srcs = glob(["*.java"]),
|
||||
assets = [
|
||||
":binary_graph",
|
||||
":model",
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/assets:box.obj.uuu",
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/assets:classic_colors.png",
|
||||
] + select({
|
||||
"//conditions:default": [
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/assets/sneaker:model.obj.uuu",
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/assets/sneaker:texture.jpg",
|
||||
],
|
||||
":use_chair_model": [
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/assets/chair:model.obj.uuu",
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/objectdetection3d/assets/chair:texture.jpg",
|
||||
],
|
||||
}),
|
||||
":mesh",
|
||||
":texture",
|
||||
MODELS_DIR + ":object_detection_ssd_mobilenetv2_oidv4_fp16.tflite",
|
||||
MODELS_DIR + ":object_detection_oidv4_labelmap.pbtxt",
|
||||
ASSETS_DIR + ":box.obj.uuu",
|
||||
ASSETS_DIR + ":classic_colors.png",
|
||||
],
|
||||
assets_dir = "",
|
||||
manifest = "//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:AndroidManifest.xml",
|
||||
manifest_values = {
|
||||
"applicationId": "com.google.mediapipe.apps.objectdetection3d",
|
||||
"appName": "Objectron",
|
||||
"mainActivity": ".MainActivity",
|
||||
"cameraFacingFront": "False",
|
||||
"binaryGraphName": "object_detection_3d.binarypb",
|
||||
"inputVideoStreamName": "input_video",
|
||||
"outputVideoStreamName": "output_video",
|
||||
"flipFramesVertically": "True",
|
||||
},
|
||||
manifest_values = select({
|
||||
"//conditions:default": generate_manifest_values("com.google.mediapipe.apps.objectdetection3d_shoe", "Shoe Objectron"),
|
||||
":use_chair": generate_manifest_values("com.google.mediapipe.apps.objectdetection3d_chair", "Chair Objectron"),
|
||||
":use_cup": generate_manifest_values("com.google.mediapipe.apps.objectdetection3d_cup", "Cup Objectron"),
|
||||
":use_camera": generate_manifest_values("com.google.mediapipe.apps.objectdetection3d_camera", "Camera Objectron"),
|
||||
":use_shoe_1stage": generate_manifest_values("com.google.mediapipe.apps.objectdetection3d_shoe_1stage", "Single Stage Shoe Objectron"),
|
||||
":use_chair_1stage": generate_manifest_values("com.google.mediapipe.apps.objectdetection3d_chair_1stage", "Single Stage Chair Objectron"),
|
||||
}),
|
||||
multidex = "native",
|
||||
deps = [
|
||||
":manifest_lib",
|
||||
":mediapipe_jni_lib",
|
||||
"//mediapipe/examples/android/src/java/com/google/mediapipe/apps/basic:basic_lib",
|
||||
"//mediapipe/framework/formats:landmark_java_proto_lite",
|
||||
|
||||
+31
-1
@@ -1,4 +1,4 @@
|
||||
// Copyright 2019 The MediaPipe Authors.
|
||||
// Copyright 2020 The MediaPipe Authors.
|
||||
//
|
||||
// Licensed under the Apache License, Version 2.0 (the "License");
|
||||
// you may not use this file except in compliance with the License.
|
||||
@@ -14,6 +14,9 @@
|
||||
|
||||
package com.google.mediapipe.apps.objectdetection3d;
|
||||
|
||||
import android.content.pm.ApplicationInfo;
|
||||
import android.content.pm.PackageManager;
|
||||
import android.content.pm.PackageManager.NameNotFoundException;
|
||||
import android.graphics.Bitmap;
|
||||
import android.graphics.BitmapFactory;
|
||||
import android.os.Bundle;
|
||||
@@ -40,10 +43,25 @@ public class MainActivity extends com.google.mediapipe.apps.basic.MainActivity {
|
||||
private Bitmap objTexture = null;
|
||||
private Bitmap boxTexture = null;
|
||||
|
||||
// ApplicationInfo for retrieving metadata defined in the manifest.
|
||||
private ApplicationInfo applicationInfo;
|
||||
|
||||
@Override
|
||||
protected void onCreate(Bundle savedInstanceState) {
|
||||
super.onCreate(savedInstanceState);
|
||||
|
||||
try {
|
||||
applicationInfo =
|
||||
getPackageManager().getApplicationInfo(getPackageName(), PackageManager.GET_META_DATA);
|
||||
} catch (NameNotFoundException e) {
|
||||
Log.e(TAG, "Cannot find application info: " + e);
|
||||
}
|
||||
|
||||
String categoryName = applicationInfo.metaData.getString("categoryName");
|
||||
float[] modelScale = parseFloatArrayFromString(
|
||||
applicationInfo.metaData.getString("modelScale"));
|
||||
float[] modelTransform = parseFloatArrayFromString(
|
||||
applicationInfo.metaData.getString("modelTransformation"));
|
||||
prepareDemoAssets();
|
||||
AndroidPacketCreator packetCreator = processor.getPacketCreator();
|
||||
Map<String, Packet> inputSidePackets = new HashMap<>();
|
||||
@@ -51,6 +69,9 @@ public class MainActivity extends com.google.mediapipe.apps.basic.MainActivity {
|
||||
inputSidePackets.put("box_asset_name", packetCreator.createString(BOX_FILE));
|
||||
inputSidePackets.put("obj_texture", packetCreator.createRgbaImageFrame(objTexture));
|
||||
inputSidePackets.put("box_texture", packetCreator.createRgbaImageFrame(boxTexture));
|
||||
inputSidePackets.put("allowed_labels", packetCreator.createString(categoryName));
|
||||
inputSidePackets.put("model_scale", packetCreator.createFloat32Array(modelScale));
|
||||
inputSidePackets.put("model_transformation", packetCreator.createFloat32Array(modelTransform));
|
||||
processor.setInputSidePackets(inputSidePackets);
|
||||
}
|
||||
|
||||
@@ -134,4 +155,13 @@ public class MainActivity extends com.google.mediapipe.apps.basic.MainActivity {
|
||||
throw new RuntimeException(e);
|
||||
}
|
||||
}
|
||||
|
||||
private static float[] parseFloatArrayFromString(String string) {
|
||||
String[] elements = string.split(",", -1);
|
||||
float[] array = new float[elements.length];
|
||||
for (int i = 0; i < elements.length; ++i) {
|
||||
array[i] = Float.parseFloat(elements[i]);
|
||||
}
|
||||
return array;
|
||||
}
|
||||
}
|
||||
|
||||
+1
-1
@@ -1,4 +1,4 @@
|
||||
# Copyright 2019 The MediaPipe Authors.
|
||||
# Copyright 2020 The MediaPipe Authors.
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
|
||||
+7
-2
@@ -11,6 +11,11 @@
|
||||
# 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.
|
||||
"""MediaPipe Python Examples."""
|
||||
|
||||
from mediapipe.examples.python.upper_body_pose_tracker import UpperBodyPoseTracker
|
||||
licenses(["notice"])
|
||||
|
||||
package(default_visibility = ["//visibility:public"])
|
||||
|
||||
exports_files(
|
||||
srcs = glob(["**"]),
|
||||
)
|
||||
BIN
Binary file not shown.
BIN
Binary file not shown.
|
After Width: | Height: | Size: 339 KiB |
+1
-1
@@ -1,4 +1,4 @@
|
||||
# Copyright 2019 The MediaPipe Authors.
|
||||
# Copyright 2020 The MediaPipe Authors.
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
|
||||
+21
@@ -0,0 +1,21 @@
|
||||
# Copyright 2020 The MediaPipe Authors.
|
||||
#
|
||||
# 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.
|
||||
|
||||
licenses(["notice"])
|
||||
|
||||
package(default_visibility = ["//visibility:public"])
|
||||
|
||||
exports_files(
|
||||
srcs = glob(["**"]),
|
||||
)
|
||||
BIN
Binary file not shown.
BIN
Binary file not shown.
|
After Width: | Height: | Size: 256 KiB |
+1
-1
@@ -1,4 +1,4 @@
|
||||
# Copyright 2019 The MediaPipe Authors.
|
||||
# Copyright 2020 The MediaPipe Authors.
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
|
||||
+14
@@ -0,0 +1,14 @@
|
||||
"""Build defs for Objectron."""
|
||||
|
||||
def generate_manifest_values(application_id, app_name):
|
||||
manifest_values = {
|
||||
"applicationId": application_id,
|
||||
"appName": app_name,
|
||||
"mainActivity": "com.google.mediapipe.apps.objectdetection3d.MainActivity",
|
||||
"cameraFacingFront": "False",
|
||||
"binaryGraphName": "object_detection_3d.binarypb",
|
||||
"inputVideoStreamName": "input_video",
|
||||
"outputVideoStreamName": "output_video",
|
||||
"flipFramesVertically": "True",
|
||||
}
|
||||
return manifest_values
|
||||
+17
@@ -0,0 +1,17 @@
|
||||
<?xml version="1.0" encoding="utf-8"?>
|
||||
<manifest xmlns:android="http://schemas.android.com/apk/res/android"
|
||||
package="com.google.mediapipe.apps.objectdetection3d">
|
||||
|
||||
<uses-sdk
|
||||
android:minSdkVersion="21"
|
||||
android:targetSdkVersion="27" />
|
||||
|
||||
<application>
|
||||
<meta-data android:name="categoryName" android:value="Camera"/>
|
||||
<meta-data android:name="modelScale" android:value="250, 250, 250"/>
|
||||
<meta-data android:name="modelTransformation" android:value="1.0, 0.0, 0.0, 0.0,
|
||||
0.0, 0.0, 1.0, 0.0,
|
||||
0.0, -1.0, 0.0, -0.0015,
|
||||
0.0, 0.0, 0.0, 1.0"/>
|
||||
</application>
|
||||
</manifest>
|
||||
+17
@@ -0,0 +1,17 @@
|
||||
<?xml version="1.0" encoding="utf-8"?>
|
||||
<manifest xmlns:android="http://schemas.android.com/apk/res/android"
|
||||
package="com.google.mediapipe.apps.objectdetection3d">
|
||||
|
||||
<uses-sdk
|
||||
android:minSdkVersion="21"
|
||||
android:targetSdkVersion="27" />
|
||||
|
||||
<application>
|
||||
<meta-data android:name="categoryName" android:value="Chair"/>
|
||||
<meta-data android:name="modelScale" android:value="0.1, 0.05, 0.1"/>
|
||||
<meta-data android:name="modelTransformation" android:value="1.0, 0.0, 0.0, 0.0,
|
||||
0.0, 1.0, 0.0, -10.0,
|
||||
0.0, 0.0, -1.0, 0.0,
|
||||
0.0, 0.0, 0.0, 1.0"/>
|
||||
</application>
|
||||
</manifest>
|
||||
+17
@@ -0,0 +1,17 @@
|
||||
<?xml version="1.0" encoding="utf-8"?>
|
||||
<manifest xmlns:android="http://schemas.android.com/apk/res/android"
|
||||
package="com.google.mediapipe.apps.objectdetection3d">
|
||||
|
||||
<uses-sdk
|
||||
android:minSdkVersion="21"
|
||||
android:targetSdkVersion="27" />
|
||||
|
||||
<application>
|
||||
<meta-data android:name="categoryName" android:value="Coffee cup,Mug"/>
|
||||
<meta-data android:name="modelScale" android:value="500, 500, 500"/>
|
||||
<meta-data android:name="modelTransformation" android:value="1.0, 0.0, 0.0, 0.0,
|
||||
0.0, 0.0, 1.0, -0.001,
|
||||
0.0, -1.0, 0.0, 0.0,
|
||||
0.0, 0.0, 0.0, 1.0"/>
|
||||
</application>
|
||||
</manifest>
|
||||
+17
@@ -0,0 +1,17 @@
|
||||
<?xml version="1.0" encoding="utf-8"?>
|
||||
<manifest xmlns:android="http://schemas.android.com/apk/res/android"
|
||||
package="com.google.mediapipe.apps.objectdetection3d">
|
||||
|
||||
<uses-sdk
|
||||
android:minSdkVersion="21"
|
||||
android:targetSdkVersion="27" />
|
||||
|
||||
<application>
|
||||
<meta-data android:name="categoryName" android:value="Footwear"/>
|
||||
<meta-data android:name="modelScale" android:value="0.25, 0.25, 0.12"/>
|
||||
<meta-data android:name="modelTransformation" android:value="1.0, 0.0, 0.0, 0.0,
|
||||
0.0, 0.0, 1.0, 0.0,
|
||||
0.0, -1.0, 0.0, 0.0,
|
||||
0.0, 0.0, 0.0, 1.0"/>
|
||||
</application>
|
||||
</manifest>
|
||||
+21
@@ -0,0 +1,21 @@
|
||||
# Copyright 2020 The MediaPipe Authors.
|
||||
#
|
||||
# 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.
|
||||
|
||||
licenses(["notice"])
|
||||
|
||||
package(default_visibility = ["//visibility:public"])
|
||||
|
||||
exports_files(
|
||||
srcs = glob(["**"]),
|
||||
)
|
||||
@@ -51,6 +51,6 @@ cc_binary(
|
||||
name = "face_detection_tpu",
|
||||
deps = [
|
||||
"//mediapipe/examples/coral:demo_run_graph_main",
|
||||
"//mediapipe/graphs/face_detection:desktop_tflite_calculators",
|
||||
"//mediapipe/graphs/face_detection:desktop_live_calculators",
|
||||
],
|
||||
)
|
||||
|
||||
@@ -18,14 +18,23 @@ licenses(["notice"])
|
||||
|
||||
package(default_visibility = ["//mediapipe/examples:__subpackages__"])
|
||||
|
||||
FACE_DETECTION_DEPS = [
|
||||
"//mediapipe/calculators/image:image_transformation_calculator",
|
||||
"//mediapipe/calculators/tflite:ssd_anchors_calculator",
|
||||
"//mediapipe/calculators/tflite:tflite_converter_calculator",
|
||||
"//mediapipe/calculators/tflite:tflite_inference_calculator",
|
||||
"//mediapipe/calculators/tflite:tflite_tensors_to_detections_calculator",
|
||||
"//mediapipe/calculators/util:detection_label_id_to_text_calculator",
|
||||
"//mediapipe/calculators/util:detection_letterbox_removal_calculator",
|
||||
"//mediapipe/calculators/util:non_max_suppression_calculator",
|
||||
]
|
||||
|
||||
mediapipe_simple_subgraph(
|
||||
name = "autoflip_face_detection_subgraph",
|
||||
graph = "face_detection_subgraph.pbtxt",
|
||||
register_as = "AutoFlipFaceDetectionSubgraph",
|
||||
visibility = ["//visibility:public"],
|
||||
deps = [
|
||||
"//mediapipe/graphs/face_detection:desktop_tflite_calculators",
|
||||
],
|
||||
deps = FACE_DETECTION_DEPS,
|
||||
)
|
||||
|
||||
mediapipe_simple_subgraph(
|
||||
@@ -33,16 +42,7 @@ mediapipe_simple_subgraph(
|
||||
graph = "front_face_detection_subgraph.pbtxt",
|
||||
register_as = "AutoFlipFrontFaceDetectionSubgraph",
|
||||
visibility = ["//visibility:public"],
|
||||
deps = [
|
||||
"//mediapipe/calculators/image:image_transformation_calculator",
|
||||
"//mediapipe/calculators/tflite:ssd_anchors_calculator",
|
||||
"//mediapipe/calculators/tflite:tflite_converter_calculator",
|
||||
"//mediapipe/calculators/tflite:tflite_inference_calculator",
|
||||
"//mediapipe/calculators/tflite:tflite_tensors_to_detections_calculator",
|
||||
"//mediapipe/calculators/util:detection_label_id_to_text_calculator",
|
||||
"//mediapipe/calculators/util:detection_letterbox_removal_calculator",
|
||||
"//mediapipe/calculators/util:non_max_suppression_calculator",
|
||||
],
|
||||
deps = FACE_DETECTION_DEPS,
|
||||
)
|
||||
|
||||
mediapipe_simple_subgraph(
|
||||
|
||||
@@ -20,7 +20,7 @@ cc_binary(
|
||||
name = "face_detection_cpu",
|
||||
deps = [
|
||||
"//mediapipe/examples/desktop:demo_run_graph_main",
|
||||
"//mediapipe/graphs/face_detection:desktop_tflite_calculators",
|
||||
"//mediapipe/graphs/face_detection:desktop_live_calculators",
|
||||
],
|
||||
)
|
||||
|
||||
@@ -29,6 +29,6 @@ cc_binary(
|
||||
name = "face_detection_gpu",
|
||||
deps = [
|
||||
"//mediapipe/examples/desktop:demo_run_graph_main_gpu",
|
||||
"//mediapipe/graphs/face_detection:mobile_calculators",
|
||||
"//mediapipe/graphs/face_detection:desktop_live_gpu_calculators",
|
||||
],
|
||||
)
|
||||
|
||||
@@ -1,42 +0,0 @@
|
||||
# Copyright 2019 The MediaPipe Authors.
|
||||
#
|
||||
# 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.
|
||||
|
||||
licenses(["notice"])
|
||||
|
||||
package(default_visibility = ["//mediapipe/examples:__subpackages__"])
|
||||
|
||||
cc_binary(
|
||||
name = "multi_hand_tracking_tflite",
|
||||
deps = [
|
||||
"//mediapipe/examples/desktop:simple_run_graph_main",
|
||||
"//mediapipe/graphs/hand_tracking:multi_hand_desktop_tflite_calculators",
|
||||
],
|
||||
)
|
||||
|
||||
cc_binary(
|
||||
name = "multi_hand_tracking_cpu",
|
||||
deps = [
|
||||
"//mediapipe/examples/desktop:demo_run_graph_main",
|
||||
"//mediapipe/graphs/hand_tracking:multi_hand_desktop_tflite_calculators",
|
||||
],
|
||||
)
|
||||
|
||||
# Linux only
|
||||
cc_binary(
|
||||
name = "multi_hand_tracking_gpu",
|
||||
deps = [
|
||||
"//mediapipe/examples/desktop:demo_run_graph_main_gpu",
|
||||
"//mediapipe/graphs/hand_tracking:multi_hand_mobile_calculators",
|
||||
],
|
||||
)
|
||||
@@ -54,9 +54,8 @@ ios_application(
|
||||
objc_library(
|
||||
name = "FaceDetectionCpuAppLibrary",
|
||||
data = [
|
||||
"//mediapipe/graphs/face_detection:mobile_cpu_binary_graph",
|
||||
"//mediapipe/models:face_detection_front.tflite",
|
||||
"//mediapipe/models:face_detection_front_labelmap.txt",
|
||||
"//mediapipe/graphs/face_detection:face_detection_mobile_cpu.binarypb",
|
||||
"//mediapipe/modules/face_detection:face_detection_front.tflite",
|
||||
],
|
||||
deps = [
|
||||
"//mediapipe/examples/ios/common:CommonMediaPipeAppLibrary",
|
||||
|
||||
@@ -9,6 +9,6 @@
|
||||
<key>GraphInputStream</key>
|
||||
<string>input_video</string>
|
||||
<key>GraphName</key>
|
||||
<string>mobile_cpu</string>
|
||||
<string>face_detection_mobile_cpu</string>
|
||||
</dict>
|
||||
</plist>
|
||||
|
||||
@@ -54,9 +54,8 @@ ios_application(
|
||||
objc_library(
|
||||
name = "FaceDetectionGpuAppLibrary",
|
||||
data = [
|
||||
"//mediapipe/graphs/face_detection:mobile_gpu_binary_graph",
|
||||
"//mediapipe/models:face_detection_front.tflite",
|
||||
"//mediapipe/models:face_detection_front_labelmap.txt",
|
||||
"//mediapipe/graphs/face_detection:face_detection_mobile_gpu.binarypb",
|
||||
"//mediapipe/modules/face_detection:face_detection_front.tflite",
|
||||
],
|
||||
deps = [
|
||||
"//mediapipe/examples/ios/common:CommonMediaPipeAppLibrary",
|
||||
|
||||
@@ -9,6 +9,6 @@
|
||||
<key>GraphInputStream</key>
|
||||
<string>input_video</string>
|
||||
<key>GraphName</key>
|
||||
<string>mobile_gpu</string>
|
||||
<string>face_detection_mobile_gpu</string>
|
||||
</dict>
|
||||
</plist>
|
||||
|
||||
@@ -34,7 +34,7 @@ alias(
|
||||
ios_application(
|
||||
name = "FaceEffectApp",
|
||||
app_icons = ["//mediapipe/examples/ios/common:AppIcon"],
|
||||
bundle_id = BUNDLE_ID_PREFIX + ".FaceMeshGpu",
|
||||
bundle_id = BUNDLE_ID_PREFIX + ".FaceEffectGpu",
|
||||
families = [
|
||||
"iphone",
|
||||
"ipad",
|
||||
|
||||
@@ -60,7 +60,7 @@ objc_library(
|
||||
"FaceMeshGpuViewController.h",
|
||||
],
|
||||
data = [
|
||||
"//mediapipe/graphs/face_mesh:face_mesh_mobile_gpu_binary_graph",
|
||||
"//mediapipe/graphs/face_mesh:face_mesh_mobile_gpu.binarypb",
|
||||
"//mediapipe/modules/face_detection:face_detection_front.tflite",
|
||||
"//mediapipe/modules/face_landmark:face_landmark.tflite",
|
||||
],
|
||||
|
||||
@@ -55,8 +55,7 @@ objc_library(
|
||||
name = "HandDetectionGpuAppLibrary",
|
||||
data = [
|
||||
"//mediapipe/graphs/hand_tracking:hand_detection_mobile_gpu_binary_graph",
|
||||
"//mediapipe/models:palm_detection.tflite",
|
||||
"//mediapipe/models:palm_detection_labelmap.txt",
|
||||
"//mediapipe/modules/palm_detection:palm_detection.tflite",
|
||||
],
|
||||
deps = [
|
||||
"//mediapipe/examples/ios/common:CommonMediaPipeAppLibrary",
|
||||
|
||||
@@ -60,11 +60,10 @@ objc_library(
|
||||
"HandTrackingViewController.h",
|
||||
],
|
||||
data = [
|
||||
"//mediapipe/graphs/hand_tracking:hand_tracking_mobile_gpu_binary_graph",
|
||||
"//mediapipe/models:hand_landmark.tflite",
|
||||
"//mediapipe/models:handedness.txt",
|
||||
"//mediapipe/models:palm_detection.tflite",
|
||||
"//mediapipe/models:palm_detection_labelmap.txt",
|
||||
"//mediapipe/graphs/hand_tracking:hand_tracking_mobile_gpu.binarypb",
|
||||
"//mediapipe/modules/hand_landmark:hand_landmark.tflite",
|
||||
"//mediapipe/modules/hand_landmark:handedness.txt",
|
||||
"//mediapipe/modules/palm_detection:palm_detection.tflite",
|
||||
],
|
||||
deps = [
|
||||
"//mediapipe/examples/ios/common:CommonMediaPipeAppLibrary",
|
||||
|
||||
@@ -17,6 +17,10 @@
|
||||
#include "mediapipe/framework/formats/landmark.pb.h"
|
||||
|
||||
static const char* kLandmarksOutputStream = "hand_landmarks";
|
||||
static const char* kNumHandsInputSidePacket = "num_hands";
|
||||
|
||||
// Max number of hands to detect/process.
|
||||
static const int kNumHands = 2;
|
||||
|
||||
@implementation HandTrackingViewController
|
||||
|
||||
@@ -25,6 +29,8 @@ static const char* kLandmarksOutputStream = "hand_landmarks";
|
||||
- (void)viewDidLoad {
|
||||
[super viewDidLoad];
|
||||
|
||||
[self.mediapipeGraph setSidePacket:(mediapipe::MakePacket<int>(kNumHands))
|
||||
named:kNumHandsInputSidePacket];
|
||||
[self.mediapipeGraph addFrameOutputStream:kLandmarksOutputStream
|
||||
outputPacketType:MPPPacketTypeRaw];
|
||||
}
|
||||
@@ -40,12 +46,16 @@ static const char* kLandmarksOutputStream = "hand_landmarks";
|
||||
NSLog(@"[TS:%lld] No hand landmarks", packet.Timestamp().Value());
|
||||
return;
|
||||
}
|
||||
const auto& landmarks = packet.Get<::mediapipe::NormalizedLandmarkList>();
|
||||
NSLog(@"[TS:%lld] Number of landmarks on hand: %d", packet.Timestamp().Value(),
|
||||
landmarks.landmark_size());
|
||||
for (int i = 0; i < landmarks.landmark_size(); ++i) {
|
||||
NSLog(@"\tLandmark[%d]: (%f, %f, %f)", i, landmarks.landmark(i).x(),
|
||||
landmarks.landmark(i).y(), landmarks.landmark(i).z());
|
||||
const auto& multiHandLandmarks = packet.Get<std::vector<::mediapipe::NormalizedLandmarkList>>();
|
||||
NSLog(@"[TS:%lld] Number of hand instances with landmarks: %lu", packet.Timestamp().Value(),
|
||||
multiHandLandmarks.size());
|
||||
for (int handIndex = 0; handIndex < multiHandLandmarks.size(); ++handIndex) {
|
||||
const auto& landmarks = multiHandLandmarks[handIndex];
|
||||
NSLog(@"\tNumber of landmarks for hand[%d]: %d", handIndex, landmarks.landmark_size());
|
||||
for (int i = 0; i < landmarks.landmark_size(); ++i) {
|
||||
NSLog(@"\t\tLandmark[%d]: (%f, %f, %f)", i, landmarks.landmark(i).x(),
|
||||
landmarks.landmark(i).y(), landmarks.landmark(i).z());
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,79 +0,0 @@
|
||||
# Copyright 2019 The MediaPipe Authors.
|
||||
#
|
||||
# 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.
|
||||
|
||||
load(
|
||||
"@build_bazel_rules_apple//apple:ios.bzl",
|
||||
"ios_application",
|
||||
)
|
||||
load(
|
||||
"//mediapipe/examples/ios:bundle_id.bzl",
|
||||
"BUNDLE_ID_PREFIX",
|
||||
"example_provisioning",
|
||||
)
|
||||
|
||||
licenses(["notice"])
|
||||
|
||||
MIN_IOS_VERSION = "10.0"
|
||||
|
||||
alias(
|
||||
name = "multihandtrackinggpu",
|
||||
actual = "MultiHandTrackingGpuApp",
|
||||
)
|
||||
|
||||
ios_application(
|
||||
name = "MultiHandTrackingGpuApp",
|
||||
app_icons = ["//mediapipe/examples/ios/common:AppIcon"],
|
||||
bundle_id = BUNDLE_ID_PREFIX + ".MultiHandTrackingGpu",
|
||||
families = [
|
||||
"iphone",
|
||||
"ipad",
|
||||
],
|
||||
infoplists = [
|
||||
"//mediapipe/examples/ios/common:Info.plist",
|
||||
"Info.plist",
|
||||
],
|
||||
minimum_os_version = MIN_IOS_VERSION,
|
||||
provisioning_profile = example_provisioning(),
|
||||
deps = [
|
||||
":MultiHandTrackingGpuAppLibrary",
|
||||
"@ios_opencv//:OpencvFramework",
|
||||
],
|
||||
)
|
||||
|
||||
objc_library(
|
||||
name = "MultiHandTrackingGpuAppLibrary",
|
||||
srcs = [
|
||||
"MultiHandTrackingViewController.mm",
|
||||
],
|
||||
hdrs = [
|
||||
"MultiHandTrackingViewController.h",
|
||||
],
|
||||
data = [
|
||||
"//mediapipe/graphs/hand_tracking:multi_hand_tracking_mobile_gpu_binary_graph",
|
||||
"//mediapipe/models:hand_landmark.tflite",
|
||||
"//mediapipe/models:handedness.txt",
|
||||
"//mediapipe/models:palm_detection.tflite",
|
||||
"//mediapipe/models:palm_detection_labelmap.txt",
|
||||
],
|
||||
deps = [
|
||||
"//mediapipe/examples/ios/common:CommonMediaPipeAppLibrary",
|
||||
] + select({
|
||||
"//mediapipe:ios_i386": [],
|
||||
"//mediapipe:ios_x86_64": [],
|
||||
"//conditions:default": [
|
||||
"//mediapipe/graphs/hand_tracking:multi_hand_mobile_calculators",
|
||||
"//mediapipe/framework/formats:landmark_cc_proto",
|
||||
],
|
||||
}),
|
||||
)
|
||||
@@ -1,16 +0,0 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
|
||||
<plist version="1.0">
|
||||
<dict>
|
||||
<key>CameraPosition</key>
|
||||
<string>front</string>
|
||||
<key>MainViewController</key>
|
||||
<string>MultiHandTrackingViewController</string>
|
||||
<key>GraphOutputStream</key>
|
||||
<string>output_video</string>
|
||||
<key>GraphInputStream</key>
|
||||
<string>input_video</string>
|
||||
<key>GraphName</key>
|
||||
<string>multi_hand_tracking_mobile_gpu</string>
|
||||
</dict>
|
||||
</plist>
|
||||
@@ -1,21 +0,0 @@
|
||||
// Copyright 2019 The MediaPipe Authors.
|
||||
//
|
||||
// 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.
|
||||
|
||||
#import <UIKit/UIKit.h>
|
||||
|
||||
#import "mediapipe/examples/ios/common/CommonViewController.h"
|
||||
|
||||
@interface MultiHandTrackingViewController : CommonViewController
|
||||
|
||||
@end
|
||||
@@ -1,57 +0,0 @@
|
||||
// Copyright 2019 The MediaPipe Authors.
|
||||
//
|
||||
// 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.
|
||||
|
||||
#import "MultiHandTrackingViewController.h"
|
||||
|
||||
#include "mediapipe/framework/formats/landmark.pb.h"
|
||||
|
||||
static const char* kLandmarksOutputStream = "multi_hand_landmarks";
|
||||
|
||||
@implementation MultiHandTrackingViewController
|
||||
|
||||
#pragma mark - UIViewController methods
|
||||
|
||||
- (void)viewDidLoad {
|
||||
[super viewDidLoad];
|
||||
|
||||
[self.mediapipeGraph addFrameOutputStream:kLandmarksOutputStream
|
||||
outputPacketType:MPPPacketTypeRaw];
|
||||
}
|
||||
|
||||
#pragma mark - MPPGraphDelegate methods
|
||||
|
||||
// Receives a raw packet from the MediaPipe graph. Invoked on a MediaPipe worker thread.
|
||||
- (void)mediapipeGraph:(MPPGraph*)graph
|
||||
didOutputPacket:(const ::mediapipe::Packet&)packet
|
||||
fromStream:(const std::string&)streamName {
|
||||
if (streamName == kLandmarksOutputStream) {
|
||||
if (packet.IsEmpty()) {
|
||||
NSLog(@"[TS:%lld] No hand landmarks", packet.Timestamp().Value());
|
||||
return;
|
||||
}
|
||||
const auto& multi_hand_landmarks = packet.Get<std::vector<::mediapipe::NormalizedLandmarkList>>();
|
||||
NSLog(@"[TS:%lld] Number of hand instances with landmarks: %lu", packet.Timestamp().Value(),
|
||||
multi_hand_landmarks.size());
|
||||
for (int hand_index = 0; hand_index < multi_hand_landmarks.size(); ++hand_index) {
|
||||
const auto& landmarks = multi_hand_landmarks[hand_index];
|
||||
NSLog(@"\tNumber of landmarks for hand[%d]: %d", hand_index, landmarks.landmark_size());
|
||||
for (int i = 0; i < landmarks.landmark_size(); ++i) {
|
||||
NSLog(@"\t\tLandmark[%d]: (%f, %f, %f)", i, landmarks.landmark(i).x(),
|
||||
landmarks.landmark(i).y(), landmarks.landmark(i).z());
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@end
|
||||
@@ -1,208 +0,0 @@
|
||||
# Copyright 2020 The MediaPipe Authors.
|
||||
#
|
||||
# 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.
|
||||
|
||||
# Lint as: python3
|
||||
"""MediaPipe upper body pose tracker.
|
||||
|
||||
MediaPipe upper body pose tracker takes an RGB image as the input and returns
|
||||
a pose landmark list and an annotated RGB image represented as a numpy ndarray.
|
||||
|
||||
Usage examples:
|
||||
pose_tracker = UpperBodyPoseTracker()
|
||||
|
||||
pose_landmarks, _ = pose_tracker.run(
|
||||
input_file='/tmp/input.png',
|
||||
output_file='/tmp/output.png')
|
||||
|
||||
input_image = cv2.imread('/tmp/input.png')[:, :, ::-1]
|
||||
pose_landmarks, annotated_image = pose_tracker.run(input_image)
|
||||
|
||||
pose_tracker.run_live()
|
||||
|
||||
pose_tracker.close()
|
||||
"""
|
||||
|
||||
import os
|
||||
import time
|
||||
from typing import Tuple, Union
|
||||
|
||||
import cv2
|
||||
import mediapipe.python as mp
|
||||
import numpy as np
|
||||
# resources dependency
|
||||
from mediapipe.framework.formats import landmark_pb2
|
||||
|
||||
# Input and output stream names.
|
||||
INPUT_VIDEO = 'input_video'
|
||||
OUTPUT_VIDEO = 'output_video'
|
||||
POSE_LANDMARKS = 'pose_landmarks'
|
||||
|
||||
|
||||
class UpperBodyPoseTracker:
|
||||
"""MediaPipe upper body pose tracker."""
|
||||
|
||||
def __init__(self):
|
||||
"""The init method of MediaPipe upper body pose tracker.
|
||||
|
||||
The method reads the upper body pose tracking cpu binary graph and
|
||||
initializes a CalculatorGraph from it. The output packets of pose_landmarks
|
||||
and output_video output streams will be observed by callbacks. The graph
|
||||
will be started at the end of this method, waiting for input packets.
|
||||
"""
|
||||
# MediaPipe package root path
|
||||
root_path = os.sep.join( os.path.abspath(__file__).split(os.sep)[:-4])
|
||||
mp.resource_util.set_resource_dir(root_path)
|
||||
|
||||
self._graph = mp.CalculatorGraph(
|
||||
binary_graph_path=os.path.join(
|
||||
root_path,
|
||||
'mediapipe/graphs/pose_tracking/upper_body_pose_tracking_cpu.binarypb'
|
||||
))
|
||||
self._outputs = {}
|
||||
for stream_name in [POSE_LANDMARKS, OUTPUT_VIDEO]:
|
||||
self._graph.observe_output_stream(stream_name, self._assign_packet)
|
||||
self._graph.start_run()
|
||||
|
||||
def run(
|
||||
self,
|
||||
input_frame: np.ndarray = None,
|
||||
*,
|
||||
input_file: str = None,
|
||||
output_file: str = None
|
||||
) -> Tuple[Union[None, landmark_pb2.NormalizedLandmarkList], np.ndarray]:
|
||||
"""The run method of MediaPipe upper body pose tracker.
|
||||
|
||||
MediaPipe upper body pose tracker takes either the path to an image file or
|
||||
an RGB image represented as a numpy ndarray and it returns the pose
|
||||
landmarks list and the annotated RGB image represented as a numpy ndarray.
|
||||
|
||||
Args:
|
||||
input_frame: An RGB image represented as a numpy ndarray.
|
||||
input_file: The path to an image file.
|
||||
output_file: The file path that the annotated image will be saved into.
|
||||
|
||||
Returns:
|
||||
pose_landmarks: The pose landmarks list.
|
||||
annotated_image: The image with pose landmarks annotations.
|
||||
|
||||
Raises:
|
||||
RuntimeError: If the input frame doesn't contain 3 channels (RGB format)
|
||||
or the input arg is not correctly provided.
|
||||
|
||||
Examples
|
||||
pose_tracker = UpperBodyPoseTracker()
|
||||
pose_landmarks, _ = pose_tracker.run(
|
||||
input_file='/tmp/input.png',
|
||||
output_file='/tmp/output.png')
|
||||
|
||||
# Read an image and convert the BGR image to RGB.
|
||||
input_image = cv2.cvtColor(cv2.imread('/tmp/input.png'), COLOR_BGR2RGB)
|
||||
pose_landmarks, annotated_image = pose_tracker.run(input_image)
|
||||
pose_tracker.close()
|
||||
"""
|
||||
if input_file is None and input_frame is None:
|
||||
raise RuntimeError(
|
||||
'Must provide either a path to an image file or an RGB image represented as a numpy.ndarray.'
|
||||
)
|
||||
|
||||
if input_file:
|
||||
if input_frame is not None:
|
||||
raise RuntimeError(
|
||||
'Must only provide either \'input_file\' or \'input_frame\'.')
|
||||
else:
|
||||
input_frame = cv2.imread(input_file)[:, :, ::-1]
|
||||
|
||||
pose_landmarks, annotated_image = self._run_graph(input_frame)
|
||||
if output_file:
|
||||
cv2.imwrite(output_file, annotated_image[:, :, ::-1])
|
||||
return pose_landmarks, annotated_image
|
||||
|
||||
def run_live(self) -> None:
|
||||
"""Run MediaPipe upper body pose tracker with live camera input.
|
||||
|
||||
The method will be self-terminated after 30 seconds. If you need to
|
||||
terminate it earlier, press the Esc key to stop the run manually. Note that
|
||||
you need to select the output image window rather than the terminal window
|
||||
first and then press the key.
|
||||
|
||||
Examples:
|
||||
pose_tracker = UpperBodyPoseTracker()
|
||||
pose_tracker.run_live()
|
||||
pose_tracker.close()
|
||||
"""
|
||||
cap = cv2.VideoCapture(0)
|
||||
start_time = time.time()
|
||||
print(
|
||||
'Press Esc within the output image window to stop the run, or let it '
|
||||
'self terminate after 30 seconds.')
|
||||
while cap.isOpened() and time.time() - start_time < 30:
|
||||
success, input_frame = cap.read()
|
||||
if not success:
|
||||
break
|
||||
input_frame = cv2.cvtColor(cv2.flip(input_frame, 1), cv2.COLOR_BGR2RGB)
|
||||
input_frame.flags.writeable = False
|
||||
_, output_frame = self._run_graph(input_frame)
|
||||
cv2.imshow('MediaPipe upper body pose tracker',
|
||||
cv2.cvtColor(output_frame, cv2.COLOR_RGB2BGR))
|
||||
if cv2.waitKey(5) & 0xFF == 27:
|
||||
break
|
||||
cap.release()
|
||||
cv2.destroyAllWindows()
|
||||
|
||||
def close(self) -> None:
|
||||
self._graph.close()
|
||||
self._graph = None
|
||||
self._outputs = None
|
||||
|
||||
def _run_graph(
|
||||
self,
|
||||
input_frame: np.ndarray = None,
|
||||
) -> Tuple[Union[None, landmark_pb2.NormalizedLandmarkList], np.ndarray]:
|
||||
"""The internal run graph method.
|
||||
|
||||
Args:
|
||||
input_frame: An RGB image represented as a numpy ndarray.
|
||||
|
||||
Returns:
|
||||
pose_landmarks: The pose landmarks list.
|
||||
annotated_image: The image with pose landmarks annotations.
|
||||
|
||||
Raises:
|
||||
RuntimeError: If the input frame doesn't contain 3 channels representing
|
||||
RGB.
|
||||
"""
|
||||
|
||||
if input_frame.shape[2] != 3:
|
||||
raise RuntimeError('input frame must have 3 channels.')
|
||||
|
||||
self._outputs.clear()
|
||||
start_time = time.time()
|
||||
self._graph.add_packet_to_input_stream(
|
||||
stream=INPUT_VIDEO,
|
||||
packet=mp.packet_creator.create_image_frame(
|
||||
image_format=mp.ImageFormat.SRGB, data=input_frame),
|
||||
timestamp=mp.Timestamp.from_seconds(start_time))
|
||||
self._graph.wait_until_idle()
|
||||
|
||||
pose_landmarks = None
|
||||
if POSE_LANDMARKS in self._outputs:
|
||||
pose_landmarks = mp.packet_getter.get_proto(self._outputs[POSE_LANDMARKS])
|
||||
annotated_image = mp.packet_getter.get_image_frame(
|
||||
self._outputs[OUTPUT_VIDEO]).numpy_view()
|
||||
print('UpperBodyPoseTracker.Run() took',
|
||||
time.time() - start_time, 'seconds')
|
||||
return pose_landmarks, annotated_image
|
||||
|
||||
def _assign_packet(self, stream_name: str, packet: mp.Packet) -> None:
|
||||
self._outputs[stream_name] = packet
|
||||
Reference in New Issue
Block a user