haruiz / cvstudio

A graphical annotation tool to address different Computer Vision tasks.

Hello world, I'm Henry 👋

• Research Scientist at Texas A&M AgriLife Research
• Google Developer Expert (GDE) in Machine Learning.
• Former Microsoft Student Partner.
• Full-stack developer in Python and Javascript.
• Mobile application developer in Android using Kotlin and Java.
• Cloud platforms: GCP, Azure, and AWS.

Proficient in .Net, MATLAB, and C++. I also have experience working with Ml/Dl/Data Science libraries such as Tensorflow, PyTorch, Scikit-Learn, Scipy, Dask, Pandas, OpenVINO, and Opencv.

My journey 🚀

• 🚜 I am currently working as a Research Scientist for Texas A&M AgriLife Research. My research focuses on addressing agricultural challenges using the intersection between Data Science, Remote Sensing, Machine Learning, and Deep Learning techniques.

• 🔭 My areas of interest include:

  • The application of Geophysics tools like Ground Penetrating Radar
  • Signal processing and computer vision.
  • Machine Learning and Deep Learning.
  • Remote sensing (UAV and satellite imagery processing).
  • Geophysics data science.
  • Mathematical and electromagnetism simulation.

• 🌱 I’m currently learning Jax and Transformers in my free time.

• 👯 Looking to collaborate in any open-source project related to the research areas mentioned above or with social meaning.

• 🚀 Open-source projects: FalconCV, CVStudio and Rigging.js.

• ⚡ Fun fact: Before I went to school, I wanted to be a soccer player.

🏆 Recent achievements & Certificates

• Graduate Remote Sensing Certificate, College of Geoscienses - Texas A&M University
• Deep Learning Udacity Nanodegree
• Intel® Edge AI for IoT Developers Nanodegree
• Intel Edge AI Certification
• Tensorflow Developer Certificated

Github Activity

🗂️ Highlight Projects

📫 Connect with me

cvstudio's Issues

PyQt version

So i checked the version using pip list and gave me 5.13.0 but when i ran
from PyQt5.Qt import PYQT_VERSION_STR
print("PyQt version:", PYQT_VERSION_STR)
i got an import error message that says
builtins.ImportError: cannot import name 'PYQT_VERSION_STR' from 'PyQt5.Qt' (C:\Users\zbook\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.7_qbz5n2kfra8p0\LocalCache\local-packages\Python37\site-packages\PyQt5\Qt.pyd)

Add Toolbox Tooltips

Improve documentation

Create docker image

Update the Wiki

Train custom detection model

Deploy model to Google Cloud

testing on linux

Hi, I was trying out this new app (Thanks!) on a Linux environment, using the system python that ships with it.

I only modified the pip_req.txt by removing the forced versions for some of the dependencies (so to don't replace libraries I already have installed) I then make sure I had the other dependencies (torch etc.) were installed correctly.

My pipreq.txt looks like:

certifi==2019.6.16
cffi==1.12.3
dask
hurry.filesize==0.9
imutils==0.5.3
more-itertools==7.2.0
numpy==1.17.2
olefile==0.46
opencv-contrib-python
peewee==3.10.0
Pillow
pycparser==2.19
pyface==6.1.2
PyQt5
PyQt5
#PyQt5Designer
six==1.12.0
toolz==0.10.0
traits==5.1.2
wincertstore==0.2
marshmallow==3.2.0

Note I didn't find a package for PyQt5Designer

epinux@epinux:~$ pip3 install PyQt5Designer
ERROR: Could not find a version that satisfies the requirement PyQt5Designer (from versions: none)
ERROR: No matching distribution found for PyQt5Designer
epinux@epinux:~$ pip3 install PyQt5Designer==5.10.1
ERROR: Could not find a version that satisfies the requirement PyQt5Designer==5.10.1 (from versions: none)
ERROR: No matching distribution found for PyQt5Designer==5.10.1
epinux@epinux:~$ 

Tring to run the app I noticed some differences from the video, in my case there are no "menu-actions" in the lower-left side of the application toolbar, and I was not able to understand how load images into the app. Here a quick record of my desktop trying to run the app: https://epinux.com/index.php/s/5TTQSDw6Y79DHYK

Do you have any hints on how to debug this issue?

I don't know if useful, but from trying to click on some of the links below the dataset folder icon, I got the following error log:

python3 cvstudio.py 
libpng warning: iCCP: known incorrect sRGB profile
AttributeError: __getattribute__

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/tab_datasets.py", line 177, in open_dataset_action_slot
    tab_widget = MediaTabWidget(vo)
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/tab_media.py", line 21, in __init__
    self.media_grid=Gallery()
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/gallery/gallery.py", line 119, in __init__
    self.setupUi(self)
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/gallery/base_gallery.py", line 221, in setupUi
    QtCore.QMetaObject.connectSlotsByName(Gallery)
SystemError: <built-in function connectSlotsByName> returned a result with an error set
AttributeError: __getattribute__

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/tab_datasets.py", line 177, in open_dataset_action_slot
    tab_widget = MediaTabWidget(vo)
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/tab_media.py", line 21, in __init__
    self.media_grid=Gallery()
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/gallery/gallery.py", line 119, in __init__
    self.setupUi(self)
  File "/home/epinux/PycharmProjects/CvStudio/view/widgets/gallery/base_gallery.py", line 221, in setupUi
    QtCore.QMetaObject.connectSlotsByName(Gallery)
SystemError: <built-in function connectSlotsByName> returned a result with an error set
libpng warning: iCCP: known incorrect sRGB profile
libpng warning: iCCP: known incorrect sRGB profile
libpng warning: iCCP: known incorrect sRGB profile

Train custom classification model

Export model to Tflite

Export model to OpenVINO

Create installer

Update Wiki

cudatoolkit v10.0 conflicts with torchvision

When installing cudatoolkit, the following error occurs:

**Package cudatoolkit conflicts for:
torchvision -> pytorch[version='1.3.0.dev20191010,1.4.0.dev20191013,1.4.0.dev20191019,1.4.0.dev20191022,1.4.0.dev20191023,1.4.0.dev20191026,1.4.0.dev20191028'] -> cudatoolkit[version='>=10.1,<10.2']
cudatoolkit=10.0
pytorch -> cudatoolkit[version='>=10.1,<10.2']**

By following the pytorch installation instructions, this is easily resolved by increasing the cudatoolkit version to 10.1:
conda install pytorch torchvision cudatoolkit=10.1 -c pytorch-nightly
resolves the issue.

readme.md is just slightly out of date.

Train segmentation model

Allow users tag videos

Add Image Downloader

Code refactor

Export detections