PRD: dataset download and manifests¶
Purpose¶
The dataset step prepares the aligned CelebA face images and the metadata needed by every later experiment. The output is not just a folder of images. It is a set of manifest files that tell the other scripts where each image is, which identity it belongs to, and which experiment split should use it.
Goals¶
- Download or reuse the full CelebA aligned image dataset.
- Add the CelebA identity annotation when the image package does not include it.
- Generate full dataset manifests for images, identities, splits, and quality tags.
- Make sure every usable image has a clear destination.
- Keep debug subsets separate from the full experiment.
Inputs¶
data/raw/celeba/img_align_celeba/img_align_celeba/data/raw/celeba/identity_CelebA.txt- optional CelebA files such as attributes, bounding boxes, landmarks, and original split files
Outputs¶
data/manifests/images.csvdata/manifests/identities.csvdata/manifests/splits.csvdata/manifests/quality_tags.csv
The generated CSV files are ignored by git because they are local experiment outputs.
Scripts¶
The dataset scripts live under data/scripts/:
download_celeba.pydownloads or copies the image package and can extract it.download_celeba_identity.pydownloads the identity annotation and normalizes it toidentity_CelebA.txt.prepare_celeba_manifests.pyscans the full dataset and writes the manifest CSV files.
The scripts are files on disk. They are not one line python -c commands, so the dataset setup can be reviewed and repeated.
Split meaning¶
gallerycontains the reference images used to build the identity database.query_knowncontains other images of identities already present in the gallery.query_unknowncontains identities that do not appear in the gallery.cluster_mixis available for grouping experiments when labels are not used.
The identity file is what makes gallery, query_known, and query_unknown possible. Without it, the project can still cluster embeddings, but it cannot measure known identity lookup or unknown rejection correctly.
Acceptance criteria¶
- The manifest contains the full aligned image set.
- The split summary reports gallery, known query, and unknown query counts when identity labels are available.
- The zip files used for download are removed after extraction unless debugging requires keeping them.
- Kaggle credentials never enter the repository.
- The output can be used by HOG, ArcFace, FAISS, DBSCAN, and the final report.