b1cfbe458c0bf123591348c54973d49297fd55ab,tools/manual/detected_faces.py,_DiskIO,_background_extract,#_DiskIO#Any#Any#,389

Before Change

                face.hash, b_image = encode_image_with_hash(aligned.face, extension)
                saver.save(output, b_image)
                final_faces.append(face.to_alignment())
            self._alignments.data[basename]["faces"] = final_faces
        saver.close()

After Change

                                      image=image,
                                      centering="head",
                                      size=512)  // TODO user selectable size
                meta = dict(alignments=face.to_png_meta(),
                            source=dict(alignments_version=self._alignments.version,
                                        original_filename=output,
                                        face_index=face_idx,
                                        source_filename=src_filename,
                                        source_is_video=self._globals.is_video))

                b_image = encode_image(aligned.face, extension, metadata=meta)
                saver.save(output, b_image)
        saver.close()

Instances