How rongalidhanush/svm-opticompress works
SVM-OptiCompress is an offline MATLAB image-codec and exploration application. A user opens the app, selects an image and encoder, compresses it into a .soc file, and can decode that file to reconstruct the image; the app also presents a six-image comparison and quality metrics. The documented encoder path measures image blocks, selects compression classes (optionally with selective SVM review), and writes quantized DCT coefficients; the decoder reads the stored stream. The graph also preserves documented training, analysis, and comparison capabilities. Source excerpts were unavailable, so relationships are limited to those stated in the README.
Main entry points: compress_image.m Image encoder · OptiCompressApp.m Interactive app · label_blocks.m SVM classification · run_project.m Project launcher
Components 21
Application
- Project launcher
run_project.m - Interactive app
app/OptiCompressApp.m - Six-image gallery
app/draw_six_images.m
Image codec
- Block extraction
src/preprocessing/image_blocks.m - Image encoder
src/compression/compress_image.m - DCT and quantization
src/compression/dct_basis.m - Quantization settings
src/compression/quant_steps.m - Zigzag scan
src/compression/zig_order.m - SOC compressed file
- Image decoder
src/decompression/decompress_image.m
Features and learning
- Block features
src/features/block_features.m - SVM classification
src/svm/label_blocks.m - Trained SVM model
src/svm/get_model.m - Selective review
research/selective_quantizers.m - Model training
train_svm_model.m
Analysis and comparison
- Block clustering
src/features/cluster_blocks.m - Spatial transitions
src/features/spatial_transitions.m - SVM and PCA demos
src/optimization - Method comparison
src/metrics/compare_methods.m - Image quality metrics
src/metrics/image_metrics.m
Other components
- Image user
Connections 21
- Image user → Project launcher: launches
README.md - Project launcher → Interactive app: opens
README.md - Image user → Interactive app: selects image
README.md - Interactive app → Six-image gallery: shows comparison
README.md - Interactive app → Image encoder: compresses image
README.md - Block extraction → Block features: measures
README.md - Block features → SVM classification: describes blocks
README.md - Trained SVM model → SVM classification: provides model
README.md - SVM classification → Selective review: scores settings
README.md - Selective review → Image encoder: reviews blocks
README.md - Image encoder → DCT and quantization: transforms blocks
README.md - Image encoder → Quantization settings: quantizes coefficients
README.md - Image encoder → Zigzag scan: scans coefficients
README.md - Image encoder → SOC compressed file: writes stream
README.md - SOC compressed file → Image decoder: reads stream
README.md - Model training → Trained SVM model: replaces model
README.md - Image user → Block clustering: explores clusters
README.md - Block clustering → Spatial transitions: enables inspection
README.md - Image user → SVM and PCA demos: explores optimization
README.md - Image user → Method comparison: compares encoders
README.md - Method comparison → Image quality metrics: reports quality
README.md