diff --git a/.gitignore b/.gitignore index 2979fcc6..c209df46 100644 --- a/.gitignore +++ b/.gitignore @@ -139,4 +139,22 @@ dmypy.json ## Git .DS_Store -/aibox/runs \ No newline at end of file + +# Project specific directories and files +aibox/resources/datasets/ +aibox/resources/old_code/Testing/ +aibox/resources/fruithands.csv +aibox/resources/hands_fp_raw.csv +aibox/resources/object_detector_idea.pdf +aibox/example_gpu.sh +aibox/UOS_HPC_Guide.md +flobox/experiment/analysis/Ramon/ +flobox/experiment/forms/ +flobox/experiment/plots/ +flobox/experiment/Abrechnung_VPgelder.pdf +flobox/experiment/Data_Analysis.zip +flobox/resources/Copenhagen/ +flobox/resources/OptiVisT_Retreat_Poster.pdf +flobox/resources/OptiVisT_Workshop_Program.pdf +flobox/resources/OptiVisT_Workshop_Timetable.pdf +flobox/Powell2024-Helping_Blind_People_Grasp.pdf diff --git a/README.md b/README.md index a0f8249e..8f34e03a 100644 --- a/README.md +++ b/README.md @@ -38,3 +38,24 @@ You do not need all the YOLOv5 files, just the following: - export.py If you clone this repo, YOLOv5 is already set up properly and you only need to separately install the YOLOv8 packages into your environment. + + +### Exporting models for use with the NLInterface Android App + +These models can be used in the Android app developed during the study project "Making blind people grasp" in 2023/2024 lead by Peter König. +The script for that is located in `aibox/app_export.py` and works for the hand detection as well as object detection. + +To export the hand detection model, first export it as a tensorflow saved model using the default export script. + +```bash +$ cd aibox +$ python export.py --weights hand.pt --include saved_model --keras +``` +This will create the `hand_saved_model` directory. + +After that you can convert it for the Tensorflow-Lite Runtime using + +```bash +$ python app_export.py --hands hand_saved_model --output hands.tflite +``` +To run the master.py file, change directory into the aibox folder before executing the code (pathing issue). \ No newline at end of file diff --git a/Shape_detection/Images/0.jpg b/Shape_detection/Images/0.jpg new file mode 100644 index 00000000..b5fe20fd Binary files /dev/null and b/Shape_detection/Images/0.jpg differ diff --git a/Shape_detection/Images/1.jpg b/Shape_detection/Images/1.jpg new file mode 100644 index 00000000..dc0e7b71 Binary files /dev/null and b/Shape_detection/Images/1.jpg differ diff --git a/Shape_detection/Images/2.jpg b/Shape_detection/Images/2.jpg new file mode 100644 index 00000000..36157ce6 Binary files /dev/null and b/Shape_detection/Images/2.jpg differ diff --git a/Shape_detection/Images/3.jpg b/Shape_detection/Images/3.jpg new file mode 100644 index 00000000..6e3600bb Binary files /dev/null and b/Shape_detection/Images/3.jpg differ diff --git a/Shape_detection/Images/4.jpg b/Shape_detection/Images/4.jpg new file mode 100644 index 00000000..20f8e949 Binary files /dev/null and b/Shape_detection/Images/4.jpg differ diff --git a/Shape_detection/Images/5.jpg b/Shape_detection/Images/5.jpg new file mode 100644 index 00000000..56f8a78c Binary files /dev/null and b/Shape_detection/Images/5.jpg differ diff --git a/Shape_detection/Images/6.jpg b/Shape_detection/Images/6.jpg new file mode 100644 index 00000000..eaeb27e5 Binary files /dev/null and b/Shape_detection/Images/6.jpg differ diff --git a/Shape_detection/Images/7.jpg b/Shape_detection/Images/7.jpg new file mode 100644 index 00000000..45cf8e84 Binary files /dev/null and b/Shape_detection/Images/7.jpg differ diff --git a/Shape_detection/Images/8.jpg b/Shape_detection/Images/8.jpg new file mode 100644 index 00000000..7a2522dc Binary files /dev/null and b/Shape_detection/Images/8.jpg differ diff --git a/Shape_detection/Images/9.jpg b/Shape_detection/Images/9.jpg new file mode 100644 index 00000000..f2adea44 Binary files /dev/null and b/Shape_detection/Images/9.jpg differ diff --git a/Shape_detection/Images/a.jpg b/Shape_detection/Images/a.jpg new file mode 100644 index 00000000..cd0d53ee Binary files /dev/null and b/Shape_detection/Images/a.jpg differ diff --git a/Shape_detection/Images/b.jpg b/Shape_detection/Images/b.jpg new file mode 100644 index 00000000..d3170488 Binary files /dev/null and b/Shape_detection/Images/b.jpg differ diff --git a/Shape_detection/Images/c.jpg b/Shape_detection/Images/c.jpg new file mode 100644 index 00000000..48753bf3 Binary files /dev/null and b/Shape_detection/Images/c.jpg differ diff --git a/Shape_detection/Images/d.jpg b/Shape_detection/Images/d.jpg new file mode 100644 index 00000000..dbd30a82 Binary files /dev/null and b/Shape_detection/Images/d.jpg differ diff --git a/Shape_detection/Images/e.jpg b/Shape_detection/Images/e.jpg new file mode 100644 index 00000000..85a7a25a Binary files /dev/null and b/Shape_detection/Images/e.jpg differ diff --git a/Shape_detection/Images/emnist.zip b/Shape_detection/Images/emnist.zip new file mode 100644 index 00000000..0e514d36 --- /dev/null +++ b/Shape_detection/Images/emnist.zip @@ -0,0 +1,1458 @@ + + +
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +The Information Technology Laboratory (ITL) is one of NIST’s six research laboratories. ITL focuses on IT measurements, testing, and standards, and is a globally recognized and trusted source of high-quality, independent, and unbiased research and data. ITL’s mission, to cultivate trust in information technology (IT) and metrology, is accomplished using its world-class measurement and testing facilities and encompassing a wide range of areas of computer science, mathematics, statistics, and systems engineering.
+ +This non-regulatory role, along with ITL’s deep technical expertise in all fields of information technology, helps increase trust in IT worldwide.
+ +Published bimonthly, the ITL Newsletter features news articles on ITL's programs, projects, activities, and accomplishments. It announces selected new ITL publications and describes upcoming technical conferences and workshops.
+ + +The ITL Bulletin focuses on ITL’s research and collaborative activities in cybersecurity. It often announces new ITL publications in information security and gives an overview of the research contained within.
+ + +