DisplayAnalysis

Challenge: Quantify subtle display defects like temporal flicker and spatial non-uniformity from video recordings—subjective visual assessment lacks precision and reproducibility. Approach: Designed modular metric computation pipeline with frame-by-frame processing. ROI (Region of Interest) handling parses coordinates, crops frames to analysis zones. Per-channel (RGB) metrics include: (1) Flicker calculation via FFT (Fast Fourier Transform) detecting temporal frequency anomalies, (2) Channel instability measuring frame-to-frame variance, (3) Spatial uniformity quantifying brightness consistency across regions, (4) Texture analysis via local variance. WorstCase state tracking optimizes long video analysis by sampling frames and maintaining peak metrics rather than processing every frame. FPS detection ensures accurate temporal analysis. CLI interface (argparse) supports batch processing; dual-mode tkinter GUI provides real-time preview with threading/queue architecture preventing UI freezes. PDF report generation includes matplotlib visualizations of metric trends and histograms. Comprehensive pytest suite validates algorithms with mocked frames, edge cases (single-frame videos, extreme values), and FPS detection accuracy. Dockerfiles separate CLI/GUI deployments. Innovation: Signal processing approach (FFT for flicker) brings engineering rigor to subjective quality assessment.