[LDA] - (Mods) Filter mods - Users can filter photos based on metadata, rating, tags, and custom criteria.

[QuentinCraft]

Filters

Each image is managed as a bitmap array on our solution.

The goal will be to apply some sorting functions. The final feature will allows the user to select any of these algorithms :

He will be able to select an category filter that allows to use all of the attached algorithms without having to select one by one.

Image Quality

• Sharpness: Use edge detection algorithms (like Sobel or Laplacian filters) to assess sharpness.
• Focus Measure: Compute the variance of the Laplacian, which measures focus clarity.
• Resolution: Select the bitmap with the largest dimensions (width × height).
• Compression Artifacts: Check for noise or artifacts using structural similarity index (SSIM).

Content Similarity

• Histogram Comparison: Compare histograms of images to find the one closest to the reference.
• Feature Matching: Use algorithms like SIFT, SURF, or ORB to compare key features between images.
• Template Matching: For bitmaps containing patterns or logos, use template matching with correlation coefficients.

Color or Visual Aesthetics

• Average Color: Select the image with the most vibrant or neutral color palette.
• Color Harmony: Evaluate color composition based on aesthetic color theories.
• Brightness and Contrast: Use statistical measures to find well-lit and balanced images.

Metadata

• Date and Time: Select the most recent or oldest bitmap.

Perceptual Metrics

• SSIM (Structural Similarity Index): Measures similarity between two images, focusing on structural content.
• PSNR (Peak Signal-to-Noise Ratio): Assesses image quality relative to a reference.
• Perceptual Hashing: Generate a hash of each image and compare for similarity.

[QuentinCraft]

Task selected

Now, I have started to implement the creation of the filter base module, that will welcome all these mechanisms and algorithms.