Searching for L/T Dwarf Candidates in Pan-STARRS Survey Images Using YOLO Object Detection

This dataset is produced from a search for L/T dwarf candidates using a deep-learning object-detection framework applied to five-band Pan-STARRS FITS images. The analysis region is located in the overlap area between the Pan-STARRS and DES surveys. After data cleaning and candidate detection with the YOLOv13 model, the detected sources were further processed through catalog cross-matching, multi-survey photometric completion, photo-type classification, LePhare validation, and proper-motion inspection. As a result, a relatively reliable catalog of known and candidate L/T dwarfs was obtained for this sub-area.

The uploaded data include two CSV files. One file contains a compiled sample of 312 known L/T dwarfs and L/T dwarf candidates used in the analysis, representing the main reference and validated candidate set in this work. The other file contains 815 image-detected candidates that could not be matched with sufficient photometric measurements in DES, VHS/VIKING, and AllWISE, and were therefore excluded from the subsequent photometric classification and validation steps. These objects remain useful as a supplementary candidate pool for future photometric completion, motion verification, or spectroscopic follow-up.

PHOTO_TYPE:Numeric value representing the photometric spectral type ranging L0 to T9, with L0 = 10 and T9 = 29