Skip to content

Low-Overhead Compression-Aware Channel Filtering for Hyperspectral Image Compression

Notifications You must be signed in to change notification settings

lidq92/LoCaCF-HSIC

Repository files navigation

LoCaCF-HSIC

Requirements

git clone https://github.com/OSGeo/gdal.git # '3.11.0dev-a8117d90bf'
cd gdal && mkdir build && cd build
# sudo apt install swig
cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_FLAGS_RELEASE="-O3 -march=native -DNDEBUG" \
    -DCMAKE_CXX_FLAGS_RELEASE="-O3 -march=native -DNDEBUG" -DBUILD_PYTHON_BINDINGS=ON ..
make -j
sudo make install
# ...

Preparation: Data and Base Codecs

Data

# download ARAD 1K dataset from https://github.com/boazarad/ARAD_1K
ln -s ARAD_1K_HSI_abs_path data/ARAD

data/ARAD/ARAD_1K_09xx.tif

Base Codecs

See Codecs.py for HEVC and VVC, etc.

Encode

python encode.py -i data/ARAD/ARAD_1K_0901.tif -codec JPEG2000 -Q 6

Decode

python decode.py -i outputs/ARAD_1K_0901_JPEG2000_s1.0_Q6.0_D0_prec32/ARAD_1K_0901.bin -org data/ARAD/ARAD_1K_0901.tif

Experiments

python Codecs.py # run SOTA codecs (including preprocessing of the datasets)
python run_ARAD.py # run LoCaCF-HSIC: SOTA + LoCaCF
# Code of LineRWKV: https://github.com/diegovalsesia/linerwkv
# We provide the results of LineRWKV (+ LoCaCF) in SOTA_results
python BD_metrics.py # BD metrics and R-D curves

TODO

FAQ

Contact

Dingquan Li @ Pengcheng Laboratory, dingquanli AT pku DOT edu DOT cn.

About

Low-Overhead Compression-Aware Channel Filtering for Hyperspectral Image Compression

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages