ukitomato / mpiristracker

This is an ios framework for Iris detection based on MediaPipe

Hi there 👋

mpiristracker's Issues

你好，这个库可以对静态的图片做面部特征识别吗？可以告诉我实现步骤吗？非常感谢

你好，我想只对一张有人脸的静态图片识别面部特征点，这个怎么实现？

非常感谢

Build error

I am unable to build the MPIrisTracker. When I run it as is, this is the error:

mediapipe/examples/ios/MPIrisTracker/MPIrisTracker.mm:10:9: fatal error: 'MyCameraInputSource.h' file not found
#import "MyCameraInputSource.h"
        ^~~~~~~~~~~~~~~~~~~~~~~
1 error generated.

So I added the new .h and .m files to the BUILD instructions as follows:

objc_library(
    name = "IrisTrackerLibrary",
#     srcs = [
#         "MPIrisTracker.mm",
#     ],
# 	hdrs = [
#         "MPIrisTracker.h",
#  	],

	srcs= glob(["*.mm"]),
	hdrs = glob(["*.h"]),

But then I get a new error:

ERROR: /Users/ayaeid/Git/mediapipe/mediapipe/examples/ios/MPIrisTracker/BUILD:3:14: Linking mediapipe/examples/ios/MPIrisTracker/MPIrisTracker_bin failed: (Aborted): sandbox-exec failed: error executing command 
  (cd /private/var/tmp/_bazel_ayaeid/f877e822d5b4000695469eba8106cb8b/sandbox/darwin-sandbox/2275/execroot/mediapipe && \
  exec env - \
    APPLE_SDK_PLATFORM=iPhoneOS \
    APPLE_SDK_VERSION_OVERRIDE=16.2 \
    DEVELOPER_DIR=/Applications/Xcode.app/Contents/Developer \
    PATH=/Users/ayaeid/Library/Caches/bazelisk/downloads/bazelbuild/bazel-5.2.0-darwin-x86_64/bin:/Users/ayaeid/.pyenv/shims:/usr/local/bin:/Users/ayaeid/.pyenv/bin:/Users/ayaeid/opt/anaconda3/bin:/Users/ayaeid/opt/anaconda3/condabin:/usr/local/bin:/System/Cryptexes/App/usr/bin:/usr/bin:/bin:/usr/sbin:/sbin:/Library/Apple/usr/bin \
    SDKROOT=/Applications/Xcode.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS16.2.sdk \
    TMPDIR=/var/folders/k3/z10szbcj0ls9nbgd0z64kw2h0000gn/T/ \
    XCODE_VERSION_OVERRIDE=14.2.0.14C18 \
    ZERO_AR_DATE=1 \
  /usr/bin/sandbox-exec -f /private/var/tmp/_bazel_ayaeid/f877e822d5b4000695469eba8106cb8b/sandbox/darwin-sandbox/2275/sandbox.sb /var/tmp/_bazel_ayaeid/install/612c94294bf5bb5d743864f93801f7fe/process-wrapper '--timeout=0' '--kill_delay=15' external/local_config_cc/wrapped_clang_pp @bazel-out/ios-arm64-min10-applebin_ios-ios_arm64-opt-ST-bba38d49dfd8/bin/mediapipe/examples/ios/MPIrisTracker/MPIrisTracker_bin-2.params)
Undefined symbols for architecture arm64:
  "_OBJC_CLASS_$_MyCameraInputSource", referenced from:
      objc-class-ref in libIrisTrackerLibrary.a(MPIrisTracker.o)
ld: symbol(s) not found for architecture arm64
clang++: error: linker command failed with exit code 1 (use -v to see invocation)
Error in child process '/usr/bin/xcrun'. 1

I am able to build the HelloWorld library into a .ipa file, but can't get this framework to complete. Any suggestions?

MacOS: Ventura 13.2.1
Swift: 5
xcode: 14.2
python: 3.7.16
Bazel: 6.1.1
mediapipe: checked out December 6, 2022 hash fca0f58

Can you realize the static picture detection function ？

Can you realize the static picture detection function ？

i saw ur demo , cant find api use for static picture, but i really need it ,if u can help me ,im really appreciate.

Timestamp is actually the frame number??

Uncommenting line 40 of AppDelegate as shown in the image below will output the time stamp of face landmark. It appears, in the format printed, that it's not quite written in that format.

Outputting as a decimal or int instead, it's clear the timestamp is actually just the frame number.

Is there a reason these timestamps are not in microseconds as indicating in documentation?