Crate codec_eval 

codec-eval

Image codec comparison and evaluation library.

This library provides an API-first design where external crates provide encode/decode callbacks, and this library handles quality metrics, viewing conditions, and report generation.

Quick Start

use codec_eval::{EvalSession, EvalConfig, ViewingCondition, ImageData};

let config = EvalConfig::builder()
    .report_dir("./reports")
    .viewing(ViewingCondition::desktop())
    .build();

let mut session = EvalSession::new(config);

session.add_codec("my-codec", "1.0.0", Box::new(|image, request| {
    // Your encoding logic here
    Ok(encoded_bytes)
}));

let report = session.evaluate_image("test.png", image_data)?;

Modules

• error: Error types for the library
• viewing: Viewing condition modeling for perceptual metrics
• metrics: Quality metrics (DSSIM, SSIMULACRA2, Butteraugli, PSNR)
• eval: Evaluation session and report generation
• corpus: Test corpus management
• import: CSV import for third-party results
• stats: Statistical analysis and Pareto front
• [interpolation]: Quality interpolation and polynomial fitting

Re-exports

pub use corpus::Corpus;

pub use corpus::CorpusImage;

pub use corpus::ImageCategory;

pub use error::Error;

pub use error::Result;

pub use eval::assert_perception_level;

pub use eval::assert_quality;

pub use eval::evaluate_single;

pub use eval::CodecResult;

pub use eval::CorpusReport;

pub use eval::EvalConfig;

pub use eval::EvalSession;

pub use eval::ImageReport;

pub use eval::ImageData;

pub use import::CsvImporter;

pub use import::ExternalResult;

pub use metrics::MetricConfig;

pub use metrics::MetricResult;

pub use metrics::PerceptionLevel;

pub use stats::ParetoFront;

pub use stats::RDPoint;

pub use stats::Summary;

pub use viewing::ViewingCondition;

pub use viewing::REFERENCE_PPD;

pub use corpus::SparseCheckout;

pub use corpus::SparseFilter;

pub use corpus::SparseStatus;

pub use import::CsvSchema;

pub use metrics::ColorProfile;

pub use metrics::xyb_roundtrip;

pub use stats::iqr;

pub use stats::mean;

pub use stats::median;

pub use stats::percentile;

pub use stats::percentile_u32;

pub use stats::std_dev;

pub use stats::trimmed_mean;

pub use viewing::SimulationMode;

pub use viewing::SimulationParams;

Modules

corpus

Corpus management for test image collections.

decode

Utility functions for decoding images with ICC profile extraction.

error

Error types for codec-eval operations.

eval

Evaluation session and report generation.

import

CSV import for third-party encoder results.

metrics

Quality metrics for image comparison.

stats

Statistical analysis and Pareto front calculation.

viewing

Viewing condition modeling for perceptual quality assessment.